With the increased use of calcium imaging for a plethora of research purposes, the need for tools that can accurately, quickly, and efficiently process this data is needed. As a graduate student in Mark Schnitzer's lab, I built and furthered new tools and software packages to improve analysis of calcium imaging data. While there are existing reviews covering calcium imaging analysis, I decided to expand upon the table in my dissertation to include more recent and additional older algorithms, tools, and analysis pipelines. Those interested in a brief overview of calcium imaging, see the return of the brain initiative notes.

The table below is primarily focused on tools for calcium imaging, but some of them also include tools for voltage imaging, data handling for biology, and more. The goal will be to expand this to include additional imaging analysis tools going forward and I might make it more dynamic akin to my Nobel Prize Browser.

This is a part of the brain initiative notes and can also be found at neuroscience imaging analysis tools or https://github.com/bahanonu/imaging_tools.

Note: I use cell extraction to refer to algorithms that perform cell segmentation and extract neural activity traces. Any additional papers or algorithms that should be added or suggested updates to the table, leave a comment below or open an issue on the GitHub page, I want to make sure everyone’s brilliant work is acknowledged!

Table updates Below is a changelog of updates to the table, such as adding new methods.

• 2021.11.14 - Added FIBSI, SpecSeg, FIOLA, and PatchWarp.

• 2021.11.29 - Added MVG-CNN.

Readers can leave comments and ask questions on the bahanonu.com associated blog post.

Large-scale calcium (Ca²⁺) imaging, using the change in Ca²⁺ concentration inside neurons as a relatively direct measure of neural activity, has become an essential tool for neuroscientists to probe neural ensemble activity and identify coding principles the brain uses to process sensory information, store memories, and produce behaviors (Hamel et al., 2015). Originally, chemical fluorescent dyes such as OGB-1 allowed experimenters to measure Ca²⁺ concentration in neurons and thus neuronal activity (Paredes et al., 2008). These synthetic dyes allowed researchers to begin studying detailed intracellular Ca²⁺ dynamics along with activity of multiple neurons at once. However, the dyes had drawbacks: they did not allow long-term (weeks to months) imaging of neural activity, dye loading could be problematic in certain brain cell types (e.g. those with thick cell walls) or regions (deep brain structures), dyes would target non-specifically to various cellular compartments/organelles and would leak out of cells, and they could not be targeted in a genetically defined manner (though there is limited indication that this might be possible (Tour et al., 2007)). On the other hand, genetically encoded Ca²⁺ indicators (GECIs) have helped expand the types of experiments and questions that can be answered using Ca²⁺ imaging (Chen et al., 2013; Dana et al., 2018; Mank et al., 2008) by addressing many of the issues with the Ca²⁺ dyes: using adeno-associated and other viruses, they can be expressed chronically from weeks to years ((Ziv et al., 2013) and personal experiments/observations); specific promoters upstream of the GECI gene can lead to expression in specific cell types; tagging the GECI with certain amino acid sequences can enable targeting to specific regions of the cell (e.g. the axon terminal (Broussard et al., 2018; Dreosti et al., 2009)); and many other advantages. However, once experimenters had collected their Ca²⁺ imaging movies, they faced a formidable challenge in extracting both cell locations/shapes and their associated fluorescence activity traces from each of these movies.

One method of obtaining cell shapes and activity traces is via manual identification and selection of regions-of-interest (ROI) by humans and then taking the average pixel value within said ROI on each frame to reconstruct the activity trace for that cell (ROI) during the entire movie. The manual identification can take a lot of time, requires humans to browse through the entire movie frame-by-frame, and can lead to false negatives especially for low signal-to-noise ratio cells. The later is especially true for one-photon movies and modern GECIs where the baseline is too low to allow cells to be seen except for when they are active. Beyond missing cells, the ROI approach also tends to introduce crosstalk by contaminating each cell’s activity trace with fluorescent signals from nearby or overlapping cells (while cells themselves do not physically overlap, this occurs due to optical limitations and the fact that many Ca²⁺ imaging datasets are 2D, which leads to perceived spatial overlap when a 3D volume is flattened), background (e.g. neuropil from axons and dendrites), blood vessels, and other sources. Thus, a good Ca²⁺ imaging cell extraction algorithm needs to meet several criteria to allow biologists to obtain high-quality data from their movies:

• (1) be automated—reduces time needed to process data, especially as dataset sizes grow, and removes a degree of human introduced variability from manual ROI selection;

• (2) scale well with dataset sizes—this is not required, but preferred as large-scale Ca²⁺ imaging videos are becoming more widespread;

• (3) maintain high fidelity—finding all Ca²⁺ events associated with a particular cell;

• (4) have low crosstalk—minimize number of Ca²⁺ transients or other fluorescent signals in a cell’s activity trace that are not from that cell.

To solve the above problems and meet those criteria, many groups have developed and published new methods. An early popular method used Principal Component Analysis followed by Independent Component Analysis (PCA-ICA) (Mukamel et al., 2009); this was a great step forward and met several ”good algorithm” criteria. However, it could fail in cases with a lot of correlated background noise or in regions of high cell density and activity. Over the next several years, additional techniques were published to tackle the problem and among these non-negative matrix factorization (NMF) (Pnevmatikakis et al., 2014; Maruyama et al., 2014) emerged as a promising candidate to meet many of the ”good algorithm” criteria. After publication, variants and enhancements of the method (CNMF, sc-CNMF, CNMF-E, etc.) sought to improve the cell activity traces by integrating Ca²⁺ dynamics into the model (constrained NMF—CNMF) or to compensate for issues in the original algorithm (e.g. handling of correlated background activity as seen in miniature microscope movies, see CNMF-E). In addition, there is an emerging focus on online (e.g. close as possible to real-time) processing of Ca²⁺ imaging data for closed-loop experiments and also methods to improve compression of data (Buchanan et al., 2018; Giovannucci et al., 2017).

Beyond those algorithms discussed above, a variety of other techniques have been released that either directly allow experimenters to simultaneously obtain cell images and activity traces or that focus on detecting cells or improving already acquired activity trace accuracy. For example, several publications have initial or limited methods for post-hoc automated removal of false positive signals (e.g. PCA-ICA, CNMF, etc. all output signal sources that upon manual inspection turn out to not be cells or other biologically relevant signals [e.g. dendrites]); look for this to be tackled to a greater degree going forward (I will have more on this in the future).

To help new experimenters gain a clearer picture of the landscape, see below table for a list of many different Ca²⁺ imaging cell extraction and activity trace reconstruction algorithms. This table will be updated to include any missing algorithms and further discussion about Ca²⁺ imaging analysis will take place in future posts.

A new paper from my lab has been published characterizing new voltage sensors.

• High-speed recording of neural spikes in awake mice and flies with a fluorescent voltage sensor

• A faster, brighter picture of brain cells in action - brief news article on the paper.

Note: I’m breaking up my notes from SfN into a series of posts focusing on a specific topic and mainly giving summaries of interesting posters, talks, etc. I came across.

• 457.02 - The impact of pain on motivation

  • Neil Schwarz talked about his recent science paper, see:

  • Decreased motivation during chronic pain requires long-term depression in the nucleus accumbens

• 457.05 - The anterior cingulate cortex as a substrate for chronic pain-induced depression: molecular, lesional and optogenetic evidences.

  • Observe presynaptic NMDAR independent LTP in ACC.

  • Chronic pain leads to an increase in anxiety, involvement of ACC and amygdala

  • They use a sciatic nerve compression chronic pain model and test responses in a variety of assays

    • Dark-light box test

    • Novelty suppressed feeding

    • Splash test

  • In depressed patients, ACC has decreased gray matter volume.

  • Previous studies suggest that ACC encodes aversive component of chronic pain

  • Posterior intralimbic cortex lesion recovers mechanical allodynia post chronic pain

  • Lesion ACC rescues chronic pain induced anxiodepressive behavior. Optoinhibition also rescues depressive-like behavior.

  • Fluoxetine stops change in depressive-like behavior without changes in mechanical withdrawal thresholds, suggesting a role for 5HT/serotinin.

  • Papers

    • Xu, J Neuro 2008

    • Koga, Neuron 2015

    • Liu Lab

    • Barthas 2015

• 457.06 - Nucleus Accumbens subregions dissociate encoding of values for reward and pain

  • Only a small portion of end organ injuries lead to chronic pain, want to find out what predisposes patients to develop chronic pain.

  • Note: not clear if this is a peripheral or central issue, some argue it is purely peripheral and that by fixing peripheral issues, CNS problems will disappear.

  • Connectivity between mPFC-NAcc is increased in patients who develop chronic back pain.

    • Same with white matter diffusivity

    • See Corticostriatal functional connectivity predicts transition to chronic back pain

    • New data from patients in this study indicate that mPFC-NAcc-amygdala separates chronic and sustained back pain groups.

  • Hippocampus and amygdala volume are lower in persisting back pain patients.

  • About 60% of risk of developing persistent back pain in patients is captured by white matter and connectivity between mPFC-NAcc-amygdala circuit.

  • Resting state activity, by around 29 days post SNI, large reorganization in connectivity in rodent models.

  • NAcc shell D2 amplify pain behavior in rodents.

    • 5 days post SNI, see increase in iSPN activity in slice.

    • Modulating iSPN activity with dreads causes changes in pain thresholds, e.g. increase D2 MSN activity leads to decreased thresholds

    • Implies that D2 MSNs can amplify neuropathic pain.

    • Doesn’t seem that they looked before pain to see if there is a change or used optogenetics.

  • papers        

    • Lee, J Neuro 2015

    • Mutso 2012

• 457.07 - Intracellular pathways in the Nucleus Accumbens modulate the antiallodynic actions of antidepressants.

  • See paper: RGS9-2--controlled adaptations in the striatum determine the onset of action and efficacy of antidepressants in neuropathic pain states

    • Looks at RGS9-2 genes role in TCA (Tricyclic antidepressant) responses during neuropathic pain.

    • RGS9KO mice respond earlier to TCAs in terms of mechanical allodynia and immobility

    • RGS9-2 in NAcc blocks Desipramine’s (TCA) antidepressant effects.

      • Desipramine

    • HDAC5 KO more susceptible to antidepressants.

    • Sequenced NAcc after neuropathic pain with and without RGS9 KO.

    • Genes changed in RGS9 KO are the same ones responding to Desipramine.

    • DMI treatment reverses the effect of nerve injury on gene expression.

  • Notes

    • There was a conversation afterwards regarding the effectiveness of various treatments, notes on the drugs mentioned below, will need to verify the claims.

      • Duloxetine - doesn’t work well for treating lupus induced pain

      • Pregabalin - works but with side effects

      • Cannabinoids - see Cannabinoids in Pain Management: An Update

    • If it is immune system mediated, could see effective - Apkarian

    • Why are none of the patients in Apkarian’s study improving? Are they irresponsive to treatment?

    • Do mood disorders go away fairly quickly after pain is treated?

    • Higher depression between transient vs. sustained patients? Apkarian says that neither group had a significant change in the rate/development of depression pre/post study?

• DP03 - Measurement of phasic dopamine signals in the rat nucleus accumbens core and shell in response to noxious stimuli

  • Measured dopamine concentration after noxious stimuli in rat NAcc core/shell using cyclic voltametry.

  • Core has fast, large [DA] transients after noxious stimuli while Shell has smaller, slower to reach max transients.

• 739.06 - Chronic pain causes dysfunction in reward circuitry

  • Opioid engage mesolimbic DA circuitry to produce analgesia.

  • Morphine analgesia blocked by DA antagonists or lesion of VTA-DA neurons.

  • Chronic pain is associated with reduced DA or disruption of mesolimbic circuitry.

  • Opioids fail to stimulate extracellular DA in chronic pain.

  • Cocaine, but not amphetamine, failed to stimulate DA release in chronic pain animals.

  • Suppression of DA tone in chronic pain, since cocaine blocks reuptake and depends on continuous release, would see a deficit there.

  • GABA_A receptor function is altered during chronic pain and there is a decrease in KCC2 expression while BDNF is upregulated in the VTA.

  • KCC2 function compromised in VTA GABA neurons of chronic pain animals.

  • Use CPP to study reward circuit problems during chronic pain.

  • Blocking KCC2 prevents cocaine CPP

  • Papers

    • Suppression of the morphine-induced rewarding effect in the rat with neuropathic pain: implication of the reduction in mu-opioid receptor functions in the ventral tegmental area

    • Microglia Disrupt Mesolimbic Reward Circuitry in Chronic Pain

    • Is there a common molecular pathway for addiction?

• 739.07 - Distinct subpopulations of dynorphin neurons drive aversion and reward

  • Looking at dynorphin and kappa opioid receptor (KOR) function in striatum.

  • KORs are activated in DR and VTA neurons.

  • KOR activity in dorsal NAcc produces preference but in ventral NAcc produces aversion.

  • Wireless u-iLED device for discrete spatial targeting of both dorsal and ventral NAcc in the same animal.

  • NAcc dynorphin-mediated aversion and preference is KOR dependent.

  • Ventral stimulation increased dopamine levels.

  • KOR function is enhanced during pain.

  • Decrease in PR lever pressing with CFA in mice, can rescue with NorBNI (Norbinaltorphimine, opioid antagonist).

    • Interesting that this works, not clear if they tested in SNI models

  • KOR blockade doesn’t also reverse mechanical sensitivity, suggesting that it is exclusive to the motivational aspect.

  • Stimulation of Dyn-Cre NAcc potentiated aversion after CFA.

  • Papers

    • Distinct Subpopulations of Nucleus Accumbens Dynorphin Neurons Drive Aversion and Reward

    • Inflammatory Pain Promotes Increased Opioid Self-Administration: Role of Dysregulated Ventral Tegmental Area mu Opioid Receptors

Note: I’m breaking up my notes from SfN into a series of posts focusing on a specific topic and mainly giving summaries of interesting posters, talks, etc. I came across.

There were a lot of new pain-related technologies on display this SfN, mostly with a focus on

• M16 - Optogenetic inhibition of specific populations of sensory neurons mediating bladder nociception

  • Using SNS-Ai35 mice, they show that they are able to inhibit bladder responses.

  • These are initial results, seems like it might not be super reliable, but they need higher numbers of mice to tell.

• AA35 - Two-photon imaging of light-induced nociceptive processing In vivo

  • Can image DRG neurons at some time as stimulating SNS-ChR2 in mice

  • GCaMP6m (calcium indicator) signal looks okay, DRGs only appear to respond to 10Hz stimulation.

• 

• BB9 - Epidural optic fiber implant for spinal optogenetics

  • They are able to do inhibition in the spinal cord, but have mentioned that this is likely to be challenging overall.

• DD3 - Optogenetic control of dopamine release in rodents and novel opto-dopamine probes for In vivo experiments

  • Can stimulate DA terminals and do fast scan cyclic voltammetry at the same time.

  • Takes about 1 minute for DA to recover its responsiveness, this is tested with varying intervals between light stimulation.

  • Tail pinch induces maybe a slight decrease in DA release in NAcc but see a weird oscillation in activity for several tens of seconds starting about 10 seconds after the tail pinch stops. Asked if he’s tried other pain stimuli, but didn’t indicate that they had.

• 549.02 - Optogenetic control of pain and motor circuitry

  • Outlined several designed traits in peripheral optogenetics

    • Expression over long periods of time

    • Dealing with increased movement, e.g. flexing of the spinal cord

    • Not immune privileged, e.g. don’t have BBB

    • Tissue opacity, especially in the spinal cord

  • Halorhodopsin had decreased sensitivity when placed in DRGs and need constant light illumination which might cause heating and other problems

  • To combat this, used SwiChR, which caused a 0.4 to 1.3g withthrawal threshold increase when used. Also tried iC1C2, NpHR, and YFP with no effect.

  • Increasing SST+ interneurons is aversive (at least with cFos proxy).

  • Inhibiting SST+ with hM4Di causes an increase in thresholds

  • They have tried multiple fibers along the spinal cord for optogenetics, this did not work.

  • Temperature effects on optogenetics is important.

  • AAV6-hSyn-iC++-YFP might be a useful tool for better inhibition.

  • Inhibition of muscles is possible

    • They tried a number of other variats of AAV and inhibitory channels/pumps to no effect.

  • Cannulation of mice does not affect pain thresholds, at least for Amy’s prep.

  • Papers

    • Llewekyn 2010 - Orderly recruitment of motor units under optical control in vivo

    • http://www.nature.com/nm/journal/v16/n10/full/nm.2228.html

• 549.03 - Optical control of stem cell derived motor neurons restores function to paralysed muscles

  • Trying to see whether stem cell derived MNs can be functional.

  • See that grafted MNs both are able to survive and their axons follow the old path of previous MNs’s axons to innervate muscle and form stable neuromuscular junctions (NMJs).

    • They are using embryonic MNs.

  • Using optogenetics, they are able to show graded recruitment of motor units, something that doesn’t seem possible with electrical stimulation in this situation.

  • Chronic stimulation strengthens ES-MN neuromuscular junctions, almost as if they are being exercised

  • Notes

    • https://en.wikipedia.org/wiki/Acute_coronary_syndrome

    • http://cbl.eng.cam.ac.uk/pub/Public/Wolpert/Publications/HamJonWol04.pdf

• 

• 549.04 - New paradigms in wireless light delivery

  • See recently published work: http://www.nature.com/nmeth/journal/v12/n10/full/nmeth.3536.html

  • Several others working on wireless optogenetics

    • Hirase, Riken

    • Boyden, MIT

    • Roger, U of Illinois

  • Wireless device requirements that they wanted to meet

    • Don’t need to handle animals

    • Long form experiments

    • Small device

  • Mouse as a dielectric resonator, so that the mouse enables focusing of RF field and don’t need complicated tracking. They didn’t test on larger animals, but same principle should work. She noted that small animals are dielectrical objects that support specific electromagnetic modes.

  • There is a small amount of heating (about 1C) depending on the duty cycle of the signal generator.

  • They are able to get place aversion with the wireless device and can put multiple animals together.

  • They are NOT individually addressable at the moment, they are building a silicon chip design that might allow this.

  • Multi-color and behavior triggering of the light should be allowed via a small silicon chip.

  • Designs for the devices have been made available.

  • The whole setup is silent.

  • With rats, higher order modes are excited when box is made larger, so would need to adjust.

• 

• 549.05 - An Optogenetic Demonstration of Motor Primitives in the Mouse Spinal Cord

  • Looking at muscle coordination and motor primitives.

  • Modular muscle stimulation allows probing of how different parts of spinal cord affect forces produced by muscles.

  • Showed that convergent and parallel isometric force fields are located at the dorsal and ventral spinal cord, respectively.

  • See http://www.nature.com/nrn/journal/v1/n2/fig_tab/nrn1100_101a_F3.html

  • Mapped leg responses in rodent by light stimulation in Thy1-ChR2 mouse line

  • Created a system that randomly moves the mouse’s leg around, stimulates spinal cord, and measures the direction of force. Doing this over an entire area allows one to create force maps.

  • Motor neurons have ~9ms delay and produce convergent fields.

  • ChAT excitatory interneurons produce parallel fields with ~6ms delay.

  • Thy1 force fields have a wider distribution that ChAT neurons.

  • MN and ChAT force fields are additive, e.g. there is a linear superposition of parallel and convergent fields.

• 

• 549.06 - Optogenetic control of aversive sensory circuitry

  • How is itch coded? Discussed the neural basis of itch.

  • Histamine is classically thought to be a pure sensation of itch but when placed in the deep muscle, causes pain.

  • Capsaicin is thought to be for pain, but when topical capsaicin is placed on the skin, causes itching.

  • Assumption that there is some lateral interaction that helps sharpen sensory acuity.

  • Looks at B5-I spinal interneurons neuron induced itch.

    • see http://www.ncbi.nlm.nih.gov/pubmed/24726382

  • Inhibition of B5-I neurons stops itching behavior.

  • They noticed that B5-I activation seems to silence a random interneuron.

  • Ex vivo skin prep, see http://elifesciences.org/content/4/e09674

    • apply cutaneous stimulation to the skin (including cowage, itching powder)

    • modulate interneurons in the spinal cord via ChR2

    • record from projection neurons via patch

  • 5HT can also be used as a puritagin

  • B5-I neurons inhibit itch via feedfoward inhibition

  • KOR agonist inhibits itch as well, pointing toward opioid release from B5-I neuron

  • Some B5-I neurons seem to directly inhibit PNs

  • Noticed that it is easiest to trigger APs from terminal endings

  • Light delivery isn’t the same as natural stimulation

    • Cfiber, get 4Hz stimulation response, but with opto it is variable around 2Hz.

    • Afiber, get 6Hz natural stimulation response, but with opto it is reliably at 2Hz.

  • Keratinocytes are sufficient to trigger APs if activated, can block stimuli responses by inhibiting keratinocytes.

  • Papers        

    • Unbiased classification of sensory neuron types by large-scale single-cell RNA sequencing

• 549.07 - Optogenetic dissection of visceral pain

  • Bladder pain syndrome is normally treated with short duration anesthetics.

  • Continuous vs. pulsed input causes different behavior.

  • Optogenetic modulation of distension can induce bladder pain.

  • eArchT activation in sensory neurons can inhibit bladder pain responses when bladder is artificially swelled.

  • Developing wireless LED based on previous technology

    • http://www.sciencemag.org/content/340/6129/211.figures-only

  • RF powered LED device, see Neurolux:

    • (http://www.neurolux.org/)[http://www.neurolux.org/]

  • How stable is RF field? Will this fluctuate and how does it compare to Ada Poon’s system?

    • 20 cm up from the bottom of the cage leads to only a 10% reduction in power

    • Elicits spontaneous pain and real time place preference.

• Q3 - Characterization of optogenetic activation of non-peptidergic C-fibers

  • See Ross minisymposium

  • Using ex vivo skin prep to characterize PN responses to MrgD neuron stimulation.

  • Need 2Hz, but not 0.1Hz, stimulation to induce Mrgd-Cre response to blue light stimulation.

  • Lamina I PN show responses to multple cold/hot, mechanical, and light stimulus protocols.

  • No hyperreactivity of PNs after inflammation when stimulating MrgD neurons.

I’ve written several times about dual photo-stimulation and imaging (dual photo-stimulation and imaging, freely moving dual photostimulation and imaging, dual photo-stimulation and imaging, cont’d). It looks like this SfN will have a lot of posters, talks, and excitement about this new technology. A new review in Journal of Neuroscience by several of the leading investigators in this area (Emiliani, Cohen, Deisseroth, and Häusser) gives a nice overview of several recent technologies in the area and future directions.

• All-Optical Interrogation of Neural Circuits

Haven’t uploaded previous posts in awhile, I’ll be adding the backlog of notes and papers in the coming weeks. In the meantime, the below is a good read.

• Science Isn’t Broken

Pretty interesting profile on Karl Deisseroth and the rise of optogenetics.

• Lighting The Brain - Karl Deisseroth and the optogenetics breakthrough.

I recently posted a brief note on problems with statistics that can make a paper’s study hard to reproduce (see taking care with statistics), which is a widespread problem that the biomedical sciences are starting to grapple with better. If you haven’t seen it already, NIH’s site on reproducibility is worth a look. This is a nice addition and adds to the opinion pieces, such as Begley and Ellis’s highly cited warning, and efforts at magazine to have researchers put more information about the parameters and statistics used to design and analyze a study in print to help facilitate more replicable data going forward.

• NIH program on Rigor and Reproducibility

• Drug development: Raise standards for preclinical cancer research - a highly cited paper that sounded an alarm about the reproducibility of pre-clinical research.

In addition Nature has compiled several articles and other resources into a single collection dealing with statistics for biologists.

• Statistics for biologists

update Was talking to several people. Pubmed Commons was suggested as a way to both improve reproducibility and provide an avenue for people to provide updates about a data set as they conduct more experiments. Whether people will have the inclination or time is another matter.

MathWorks has a nice summary page on dealing with large datasets in Matlab (Working with Big Data in MATLAB). This reminded me of an older article from the Sorger lab at Harvard that attempts to give a high-level view of how they manage data. It was that article along with several chats with people in my lab that convinced me to switch all imaging data from the TIFF file format to HDF going forward (has much better support for super large datasets), even if that made working with ImageJ more troublesome.

Remembering the reliance of people passing down or developing good data storage practices within the lab, I found a well-documented, standard resource outlining common practice for data storage in the sciences lacking. It would be nice to see more papers describing how people store their data and what techniques they use, e.g. CERN uses magnetic tapes for storage still vs. using hard drives. This, along with basic introductions to compression, would greatly save taxpayer and other money currently spent storing data either locally on hard drives and servers or off site on the cloud.

There are several papers out there describing efforts to standardize or share data in a particular field (e.g. fMRI data), but there doesn’t seem to be common guidelines for making this happen across neuroscience disciplines. Rather, each field (this applies to the biosciences in general) appears to have consortium or other efforts that come into vogue and either continue on through the efforts of several people or languish.

An argument could be made to create a government agency (or a group inside the NIH) whose sole purpose is to help manage and guide labs in storing data, deciding which bits (pun intended) to be shared, and the best way to describe the data (e.g. metadata) so future researchers can access and make sense of it. For example, NIDA (National Institute on Data Analysis)...but NIDA is taken by National Institute on Drug Abuse so better yet NIDSA (National Institute on Scientific Data Analysis). Some might complain that this is just adding another slow moving layer of bureaucracy, but it is sometimes shocking when you go to help other scientists analyze their data how non-standard formatting of names and other common variables is, especially by those who don’t program.

To this end, the NIH has several resources that have already implemented aspects of this vision. There is their main Data Science at NIH page, which has links to several other resources. In addition, several years back the NIH established the Working Group on Data and Informatics, which has been tasked (see NIH and Biomedical ‘Big Data’ along with DIWG's executive summary). The working group has several recommendations, some of which would be very useful if implemented broadly throughout the biomedical research community, such as a specific minimal set of metadata (and potentially specific templates for particular subfields). They also recommended to help develop quantitative training programs and identify gaps in terms of researchers trained in specific areas; though, I wonder how they will achieve the later goal.

In addition to the working group, there is also the NIH Big Data to Knowledge (BD2K) initiative, which launched in 2012 and seeks to better support digital infrastructure for managing, sharing, and utilizing big data. With the need for data scientists growing each year, efforts like this are welcome, but they appear to be largely focused on clinical or -omic (geonomic, proeteomic, etc.) datasets for the time being. Beyond the NIH, there are also other collaborative efforts moving forward, such as the Global Alliance for Genomics and Health.

In neuroscience in particular, there are projects already underway to start standardizing specific aspects of data collection and sharing, such as Neurodata Without Borders and others. There are several task forces at the International Neuroinformatics Coordinating Facility whose goal is to help develop standards for electrophysiology and neuroimaging data, but it is unclear how widespread adoption of their recommendations is or will be. Much like the NIH’s vision for data science, there will likely need to be a large push toward training young scientists early before they settle into bad habits or attempt to reinvent the wheel.

While projects like BrainFormat are great ideas and excellent to see, there is a particular problem inherent to academic research: many experiments are often single purpose/one-off affairs and the people involved might be more interested in getting a result and a paper than leaving behind an easily understood set of data for others to look through. This is a case where making data organization a part of the curriculum might accrue substantial gains down the road. This would at least ensure that people are aware of the resources available, past and current thinking, and areas for improvement. If this was a single day course, much like ethics is in many universities, this might be more easily adopted and would introduce a larger body of research scientists to tools needed to deal with the deluge of data, rather than each going off to their own lab to create ad hoc methods that suite their particular experiments.

Below are several links to articles, new stories, and resources related to formatting and standardizing big data sets.

• Working with Big Data in MATLAB

• SDCube and hybrid data storage

• Adaptive informatics for multifactorial and high-content biological data

• Considerations for developing a standard for storing electrophysiology data in HDF5

• Making data sharing work: The FCP/INDI experience

• Big data, but are we ready?

• Big data from small data: data-sharing in the `long tail' of neuroscience

• A Standard for Neuroscience Data

• Electrophysiology Task Force

• BrainFormat

• Data Sharing for Computational Neuroscience

• Trends in the production of scientific data analysis resources

• NIH Working Group on Data and Informatics NIH Big Data to Knowledge (BD2K) initiative

• Global Alliance for Genomics and Health

The idea of modulating signaling pathways in a specific manner came up during lab meeting. Optogenetics has been the leading method for precise control of neural activity either by activating or inhibiting neurons via channelrhodopsin or archaerhodopsin. Another class of optogenetic tools exist known as optoXRs that splice a light responsive element to G-protein coupled receptors. This allows one to turn on intracellular signaling pathways, e.g. through cAMP, to modulate neural activity or other functions that might not just be represented in sub-second alterations in neural spiking behavior, as is done with traditional optogenetics. Below are a couple of papers that develop and utilize several different types of optoXRs.

• Spatiotemporal Control of Opioid Signaling and Behavior - optoMOR (mu opioid receptor)

• Natural Neural Projection Dynamics Underlying Social Behavior (figure 8) - optoD1 (dopamine 1 receptor)

• Temporally precise in vivo control of intracellular signalling - opto-2AR and opto-1AR (adrenergic receptors)

As the complexity of imaging experiments grows, it is becoming necessary to model out more and more of the entire experiment to ensure that you will have enough room on the mouse/rat’s head, that you can minimize damage and suffering to the animal, and provide the best data with the given setup. To do this, I’ve been using PTC Creo (though SOLIDWORKS and Autodesk's Inventor are other alternative CAD programs) to model mouse and rat surgeries. One problem is finding a suitable model of the mouse or rat brain to use. Luckily, several groups have conducted CT scans and other imaging to help render 3D models of different rodent brains. These can be imported, cleaned, and then used to design placement of different technologies to manipulate the brain, such as optogenetic fibers. See below links for a couple resources in this area.

• The Neuroterrain 3D Mouse Brain Atlas

• Dr. Salzmann's work at Max Planck Institute for Biological Cybernetics

• Mouse Imaging Centre

• Allen Institute Brain Explorer

• Digimouse: 3D Mouse Atlas

• Cerebral cartography: a vision of its future

• The BRAIN Initiative: developing technology to catalyse neuroscience discovery

• Brain/MINDS: brain-mapping project in Japan

• The future of human cerebral cartography: a novel approach - by Markram, whose helping head Europe’s Human Brain Project.

In Michio Kaku’s excellent book giving a high-level overview of neuroscience research, The Future of the Mind, he attempts to define consciousness in a very engineering-physics perspective: there are different levels of consciousness that are defined by the number of feedback loops that the system exhibits. For example, thermostats and plants are on the low end since they only have relatively simple feedback loops. Humans are a different class of consciousness because we are able to have multiple, parallel feedback loops going on simultaneously.

Giulio Tononi and Christof Koch attempt to define a theory, called Integrated information theory , that will allow us to explain consciousness. It is an interesting read, I’ve also included a link to a differing opinion as contrast.

• Consciousness: here, there and everywhere?

• Why I Am Not An Integrated Information Theorist (or, The Unconscious Expander)

• Consciousness: Confessions of a Romantic Reductionist

• Phi: A Voyage from the Brain to the Soul

I remember reading Dias, 2014. It starts to touch a bit on Lamarckism, or the idea that characteristics acquired by the organism. In school, we are classically taught that this hypothesis was incorrect and that mutations and natural selection are what determines which traits get passed on. Epigenetic Lamarckism is becoming more popular, as is the more neutral and precise term transgenerational epigenetics.

Below are a series of papers that cover epigenetic inheritance with some notes by me.

Reviews

• Experimental evidence needed to demonstrate inter- and trans-generational effects of ancestral experiences in mammals - an interesting perspective article that helps outline and clarify several aspects of the transgenerational epigenetic debate.

• Transgenerational Epigenetic Inheritance: Myths and Mechanisms - a nice review that both defines some terms and lays out the current evidence for transgenerational inheritance in humans and other organisms.

• Epigenetic mechanisms underlying learning and the inheritance of learned behaviors

• Lamarck revisited: epigenetic inheritance of ancestral odor fear conditioning

• Molecular mechanisms for the inheritance of acquired characteristics—exosomes, microRNA shuttling, fear and stress: Lamarck resurrected?

• Environmentally induced epigenetic transgenerational inheritance of disease susceptibility

• Epigenetic memory: the Lamarckian brain

• A lingering smell? - a nice perspective on the Dias, 2014 paper.

• Understanding transgenerational epigenetic inheritance via the gametes in mammals

• Transgenerational Epigenetic Inheritance: Prevalence, Mechanisms, and Implications for the Study of Heredity and Evolution

• Bridging the transgenerational gap with epigenetic memory

• Epigenetic and epigenomic variation in Arabidopsis thaliana - this review looks at epigenetic changes in Arabidopsis thaliana, a commonly used model organisms in plant research.

• Epialleles in plant evolution

Articles

• Parental olfactory experience influences behavior and neural structure in subsequent generations - perhaps one of the stronger transgenerational epigenetic papers, mostly because they demonstrate an actual function behavioral change in offspring that would potentially be advantageous in the wild. In addition, they demonstrate a morphological change that explains these behavioral changes, namely alterations in the olfactory neuron gene expression and the size of olfactory glomeruli.

• Paternally Induced Transgenerational Environmental Reprogramming of Metabolic Gene Expression in Mammals - demonstrates that altering a parents environment (specifically causing them to be on a low protein diet) can cause alterations in their offspring’s phenotype. These changes in the offspring are correlated with alterations in DNA methylation, suggesting a mechanism for transmission.

• Ancestral dichlorodiphenyltrichloroethane (DDT) exposure promotes epigenetic transgenerational inheritance of obesity - suggest that the pesticide DDT can promote obesity in subsequent generations. While it is claimed that this is due to transgenerational epigenetic effects, it is also possible that germline mutations occurred due to exposure to the drug.

• Environmentally Induced Transgenerational Epigenetic Reprogramming of Primordial Germ Cells and the Subsequent Germ Line - investigates how agricultural fungicide vinclozolin can cause epigenetic changes, by altered DNA methylation patterns and transcriptomes, in subsequent generations.

• Epigenetic Transmission of the Impact of Early Stress Across Generations - show that chronic stress early in life, in mice, can cause DNA methylation changes and altered behavior of offspring that are reared normally.

• Epigenetic inheritance of a cocaine-resistance phenotype - show that ingestion of cocaine by parents can affect offpsring resistance to cocaine and alter prefrontal gene expression patterns.

• Paternal Stress Exposure Alters Sperm MicroRNA Content and Reprograms Offspring HPA Stress Axis Regulation - this study interestingly found that microRNAs in sperm where changed after exposure to stress and that offspring of stressed animals showed hypothalamic–pituitary–adrenal (HPA) axis (involved in stress regulation) dysregulation. It’s possible that changes in microRNA level can alter early gene expression patterns, leading to subtle changes in offspring’s later health.

• Implication of sperm RNAs in transgenerational inheritance of the effects of early trauma in mice - this paper also looks at how early trauma can lead to changes in microRNA expression.

• Mechanisms of epigenetic memory and addiction

A paper recently showed that they are able to induce reward-seeking behavior in an animal by stimulating the medial forebrain bundle (MFB), which contain ventral tegmental area fibers that release dopamine, every time a place cell fires during sleep. This would imply that the place field still encodes the same information about an animals place in the environment even during sleep and hints that it is functionally meaningful. Would have been super cool if they were able to induce aversion as well.

• Explicit memory creation during sleep demonstrates a causal role of place cells in navigation

One should compare this paper to previous false memory papers, though I like this one because it is manipulating behavior based on a single neurons activity, rather than broad spectrum reactivation.

• Optogenetic stimulation of a hippocampal engram activates fear memory recall.

• Creating a False Memory in the Hippocampus

• Bidirectional switch of the valence associated with a hippocampal contextual memory engram

• Engineering a memory with LTD and LTP

While ago I wrote briefly about effect sizes and why they might be better than p-values, see statistics: effect sizes. It seems that Basic and Applied Social Psychology has decided to ban p-values (or null hypothesis testing) from the journal.

• Psychology Journal Bans Significance Testing

• Basic and Applied Social Psychology Editorial on the null hypothesis significance testing procedure

John Ioannidis has made a career from pointing out the troubling fact that there are way too many positive results in most scientific literature than should be expected by chance. This might partially be explained by the fact that people will continue to do a study until they obtain p¡0.05 and then proceed to publish. It is also a fact that new, positive finds, whether true or not, are more likely to garner attention and publication than a null result. This is a complex issue that will probably grow to dominate scientific discussion in the coming decade.

• Why Most Published Research Findings Are False

• Power failure: why small sample size undermines the reliability of neuroscience

• Scientific method: Statistical errors

• False-Positive Psychology: Undisclosed Flexibility in Data Collection and Analysis Allows Presenting Anything as Significant

• How Many Scientists Fabricate and Falsify Research? A Systematic Review and Meta-Analysis of Survey Data

Garret Stuber has a minireview out in Nature Neuropsychopharmacology concerning miniature microscopes and fiber photometry. Similar to his paper from last year reviewing ways to image/manipulate neural activity.

• In vivo Calcium Imaging to Illuminate Neurocircuit Activity Dynamics Underlying Naturalistic Behavior

• Tools for Resolving Functional Activity and Connectivity within Intact Neural Circuits

Elizabeth M. C. Hillman came by to give a talk on some of the work in her lab, which was quite interesting. She talked a bit about SCAPE microscopy, what influenced it’s development, and how they plan on improving it. She went over some of the fundamental limitations of point scanning two-photon microscopy along with other limitations of light sheet/field microscopy. The actual demonstration of SCAPE was fairly impressive, they were able to image in 3D whole Drosophila melanogaster larvae as they moved around a dish along with volumetric imaging in mice on a ball. However, when she showed a comparison to a two-photon volumetric image taken of the same slice, the SCAPE image appears considerably worse (both in the x-y, but it appears the z resolution degrades much faster with distance). But one assumes that this can be improved. Overall, pretty impressive.

• Swept confocally-aligned planar excitation (SCAPE) microscopy for high-speed volumetric imaging of behaving organisms

She also talked about laminar optical tomography, which appeared to allow her to do 3D volumetric imaging about a decade ago. Optical tomography takes advantage of the fact that light diffuses naturally through most medium, either tissue or artificial materials. Thus, if you shine light into a tissue, some of it will emerge at a distant location. The farther away it emerge, the longer distance it traveled and one can infer that this was likely in the z direction. Thus, by measuring the quantity of light emitted at points next to the spot of tissue being illuminated by the laser, one can estimate the density of material that the light traveled through and thus begin to reconstruct the 3D structure of the tissue being imaged. See below papers for detailed discussions.

• Laminar optical tomography: demonstration of millimeter-scale depth-resolved imaging in turbid media

• Laminar optical tomography: high-resolution 3D functional imaging of superficial tissues - gives a nice overview of the technique in the methods section.

She also showed a bit of data from her wide-field imaging setup to allow her to image both oxygenated and deoxygenated hemoglobin while also doing calcium imaging (using GCaMP). While it seemed most of the work was done in young animals, there was a strike lack of correlation between blood oxygen content and neural activity. If someone could at some point do a large scale fMRI and GCaMP (or use voltage fluorescent indicators) study of cortical activity, maybe that will finally cause people to start looking at fMRI studies with more skepticism, as it relates to claiming this or that brain region is involved in a behavior or cognitive process given it lights up on fMRI scans. Below paper covers this issue in a bit more depth.

• Coupling Mechanism and Significance of the BOLD Signal: A Status Report

There are various areas in the pain field, specifically stimulus delivery, that would benefit from automation that could allow more precise dissection of neural involvement in pain response/sensation when combined with new optogenetic and other techniques, seeoptogenetics and pain, continued and optogenetics and pain.

• Puzzle Imaging: Using Large-scale Dimensionality Reduction Algorithms for Localization

While back I posted about the limits of computation. Building off the discussion about AI (advances in artificial intelligence, part 2), there is an interesting question about whether superintelligence is fundamentally limited by thermodynamics and thus can only grow so far ahead of us. I’m not sure how optical or quantum computing invalidate conclusions of the below papers, but they are interesting reads nonetheless.

• The Fundamental Physical Limits of Computation

• Ultimate physical limits to computation

• Limits on fundamental limits to computation

• Fundamental Limits to Moore's Law

• Thermodynamic Cost of Reversible Computing - see layman’s overview Computers Faster Only for 75 More Years

Allen Institute details a plethora of new mice in a recent Neuron paper.

• Transgenic Mice for Intersectional Targeting of Neural Sensors and Effectors with High Specificity and Performance

• JAX Allen Mice - mice available here.

However, I worry about GCaMP expressing mice because depending on how they are bred or what Cre line is used to drive expression, you could have an extra calcium binding protein in large numbers of cells during development and adulthood of the mice being studied. Whether this effects neural activity is unknown; though, the first place to look would be if there have been any calcineurin or other calcium binding protein overexpression studies in mice. It seems that people have performed these studies and seen changes in both cardiac function and memory. Some of the studies use modified forms of calcineurin (e.g. truncated) that may make extrapolation to constitutively expressed GCaMP problematic. However, it is worth giving pause, especially since no behavioral characterization of the mice seems to have been done in the paper.

• Restricted and Regulated Overexpression Reveals Calcineurin as a Key Component in the Transition from Short-Term to Long-Term Memory

• Genetic and Pharmacological Evidence for a Novel, Intermediate Phase of Long-Term Potentiation Suppressed by Calcineurin

• Overexpression of calcineurin in mouse causes sudden cardiac death associated with decreased density of K+ channels

• Cardiac function and electrical remodeling of the calcineurin-overexpressed transgenic mouse

• Time-dependent systolic and diastolic function in mice overexpressing calcineurin

Fabian Voigt from Fritjof Helmchen’s lab came by last month. He demonstrated several videos from their new light sheet microscopes, which illuminate a plane of a sample to allow more rapid scanning of volumes. He showed several impressive videos of the beating heart of a zebrafish to demonstrate the technique.

• High-resolution reconstruction of the beating zebrafish heart

• Rapid 3D light-sheet microscopy with a tunable lens

He also talked about electrically tunable lenses, which I ran into a couple times at Photonics West. These look super useful and allow rapid focusing without needing to move any mechanical parts, something that should both allow faster imaging of volumes while also reducing chance of system failure from wear and tear.

• Fast two-layer two-photon imaging of neuronal cell populations using an electrically tunable lens

• Tunable Optics - nice overview of tunable optics.

Was talking with Benjamin Grewe, a postdoc in the Schnitzer lab, about the recent Nature paper from Google utilizing an algorithm developed at DeepMind. DeepMind was bought by Google and it seems this allows one to go from the arXiv to Nature. Bernhard Scholkopf gives a pretty good overview of the paper and some of the related literature. I’ll build a bit on my first AI post here, advances in artificial intelligence for the first entry.

• Human-level control through deep reinforcement learning

• Playing Atari with Deep Reinforcement Learning - original arXiv paper.

• Artificial intelligence: Learning to see and act - a good layman’s overview.

This additionally got me thinking again about general artificial intelligence. Vernor Vinge and others have been predicting technological singularity or more accurately, the point at which better than human intelligence is created, for some time now. The issue still seems to be that most AI is domain specific. Even in the Atari example, that program would likely utterly fail if asked to interpret what is going on in the Oath of the Horatii or if it was asked to design a simple circuit to make lights blink when the door opens. How can narrow AI, that is AI which is very good at a specific task, be combined to make a more general AI that has order of magnitude more processing power and reasoning ability compared to a human working in the same field? Below are a number of articles exploring this issue in depth.

• The AI Revolution: The Road to Superintelligence

• Google’s Grand Plan to Make Your Brain Irrelevant

• Superintelligence: Paths, Dangers, Strategies - pretty good overview of many concepts of superintelligence, how it could arise and evolve, and potential mechanisms to ensure that it doesn’t go rouge.

• The Coming Technological Singularity: How to Survive in the Post-Human Era

• The Singularity Is Near: When Humans Transcend Biology

• The Quest for Artificial Intelligence

• We Live in a Jungle of Artificial Intelligence that will Spawn Sentience

• The AI-Box Experiment - explores the question of whether an AI can convince a human to let it out of its safety box (e.g. the box used to prevent it from directly interacting with the outside world.)

• The Basic AI Drives - interesting exploration of different drives that AI have.

In addition, this in some sense alludes to another recent post on artificial intelligence and the law. What happens when an AI system commits a crime (either civil or criminal)? Should the programmer be punished or at some point would the AI be considered mature enough to itself be punished? How would you do this? Considering it could have backed itself up, shutting it down isn’t punishment enough. How would this punishment signal be sent to other AI systems? Would they care?

For example, Palantir, BlackRock, and others offer data analysis software. It wouldn’t be far-fetched to imagine Google allowing one of DeepMind’s algorithms to take these company’s data analytics frameworks and enhance them for use in the stock markets. It could look at historical stock data, gather information on companies HQ and other infrastructure location, and correlate all of this with up-to-the-minute news via Google News. Thus, because they would have information on the types of words being used previously when a stock crashed or surged (via Google Trends), e.g. if there was an increased correlation of a company’s name in the news with negative wording, they could potentially take advantage of large fluctuations better than current high-frequency algorithmic trading. This would obviously only work for a short period of time, similar to the initial fabulous gains by some participants in the 80s and 90s when computers/rigorous statistics were first coming into play. But it could also lead to situations like the http://en.wikipedia.org/wiki/2010_Flash_Crash. Some articles exploring that event in more detail below.

• What caused the flash crash? One big, bad trade

• Findings Regarding The Market Events Of May 6, 2010

• Flash Crash: Could it happen again?

• The monolith and the markets

Manipulating circadian clock neuron firing rate resets molecular circadian rhythms and behavior -Really, the SCN is important for ? I immediately was slightly suspicious since this seemed to be common knowledge and something learned in Neuro201. Except when we learned about it, they used electrical stimulation. After a single PubMed search, I came across one of the older 1980s papers that basically already proves what this paper is demonstrating. Additional review article from the 1970s included for additional reading.

• Suprachiasmatic stimulation phase shifts rodent circadian rhythms

• The role of the suprachiasmatic nuclei in the generation of circadian rhythms in the golden hamster, Mesocricetus auratus

To this end, DrugMonkey has a rather biting post on a recent Nature paper, see Nature publishes overwhelmingly proven ``NEW AMAZING FINDING''....because optogenetics!. Anyways below is the new Nature paper and a cadre of older papers looking at the behavior investigated.

• Thirst driving and suppressing signals encoded by distinct neural populations in the brain - the new paper

• Why Do We Feel Thirst? An Interview with Yuki Oka Anteroventral Wall Of The Third Ventricle And Dorsal Lamina Terminalis: Headquarters For Control Of Body Fluid Homeostasis?

• Angiotensin, Thirst, and Sodium Appetite

• Drinking behavior following electrical stimulation of the subfornical organ in the rat. - paper from 1983 where they do a similar study only using eletrical stimulation.

• Metabotropic glutamate receptors in median preoptic neurons modulate neuronal excitability and glutamatergic and GABAergic inputs from the subfornical organ.

• Endogenous angiotensin II facilitates GABAergic neurotransmission afferent to the Na+-responsive neurons of the rat median preoptic nucleus.

• Angiotensinergic and cholinergic receptors of the subfornical organ mediate sodium intake induced by GABAergic activation of the lateral parabrachial nucleus.

• Activation of the renin-angiotensin system, specifically in the subfornical organ is sufficient to induce fluid intake.

Also, I noticed that Charles Zuker has seemingly only published in Cell/Nature/Science over the last decade (see Zuker lab publications). Didn’t know that was possible.

It is currently extremely difficult to measure electrical activity in an axon coming from a specific brain region. For these reasons, fluorescent protein indicators of neural activity are used, such as GCaMP. One can do this be either expressing GCaMP at high levels such that it diffuses down the axon or by tagging it with synaptophysin GCaMP (SyGCaMP) to help localize it to axon terminals. This later strategy has been used before and seems promising for future work, see below for a couple articles.

• Concurrent Imaging of Synaptic Vesicle Recycling and Calcium Dynamics

• In vivo evidence that retinal bipolar cells generate spikes modulated by light

• Functional imaging in the zebrafish retinotectal system using RGECO

• Parametric Functional Maps of Visual Inputs to the Tectum

• Abstract - FENS Forum 2014

• Imaging zebrafish neural circuitry from whole brain to synapse

There is a rather worrying trend toward National Institute of Health (NIH) R01 grants (one of the main funding mechanisms used by the NIH to fund research) going to older and older investigators. While this helps maintain large labs and fund a variety of important projects, there is an argument to be made that this money is not being spent in the most effective manner.

This argument can come from several angles, the most obvious is using the criteria of scientific productivity. Using Nobel Prizes as the gold standard, across biology, chemistry and physics the most productive years are between 34-38 years.(Stephan and Levin, 1993) Using other criteria, it was found that productivity of biomedical scientists appears to peak in their early 30s.(Falagas et al., 2008) However, when one looks at the age distribution of NIH R01 investigators (Fig. Figure 30), the number of faculty and new investigators getting grants in this age range is dropping fast. However, it is unlikely that this will change unless there is a revolution in the NIH’s incentives and structure. I cannot find data listing the ages of the NIH review committees, which might provide some insight into the biases that might be driving this trend.

• Age Distribution of NIH Principal Investigators and Medical School Faculty

Below are a list of several publications that look at scientific productivity as a function of age:

• Age and scientific productivity. Differences between fields of learning

• At what age do biomedical scientists do their best work?

• Age and the Nobel prize revisited

• Productivity versus age

• Fostering the Independence of New Investigators in Biomedical Research. - Where Are We Now?

• Report on Enhancing Peer Review at NIH Implementation Plan

This also gets into questions how the current state of postdoctoral research, but I will save that for another time. In the meantime, below are a couple interesting reads.

• The Evolving Postdoctoral Experience

• Between Postdoc and Job, a Whole Lot of Questions

• Postdoc payday: Salaries for fellows are on the up

• The rise of the professional master’s degree: the answer to the postdoc/PhD bubble

Last year I discussed several papers that imaged zebrafish using light sheet microscopy, see whole animal 3D imaging. Now that Misha Ahrens has his own lab at Janelia, he has advanced the work he published two years ago. He has an excellent primer in Neuron going over different methods to visualize animal neural activity in 3D.

• Visualizing Whole-Brain Activity and Development at the Single-Cell Level Using Light-Sheet Microscopy

• Do rats have a prefrontal cortex?

This also gets at another issue, which is how closely findings in lower animals, such as the widely used rodent, match similar studies in humans. There are many arguments both in favor and against animal models, from the number of pre-clinical animal model drug studies that don’t end up having the same effect in patients. There need to be more studies like Becerra, 2013 that measure experimental variables in both humans and rodents to help assuage fears that rodents neural activity is fundamentally different in response to pain.

• Analogous responses in the nucleus accumbens and cingulate cortex to pain onset (aversion) and offset (relief) in rats and humans

• The necessity of animal models in pain research

• Do Animal Models Tell Us about Human Pain?

I previously talked about spatial navigation in bats and while at a graduate student seminar by Hannah Frank in the Hadly Lab, the topic of GPS navigation in bats came up. I vaguely remembered that the Ulanovsky Lab in Israel had done similar work and it seems that is actually the case. See below for a couple articles on the subject. It would be amazing if someone could record from the bat’s hippocampus during these long flights to see how spatial information is represented on these scales, e.g. do place cells expand the size of their place fields, respond to many more places along the bats path, or does a higher-level code relating the activity of multiple place fields kick in?

• Spatial cognition in bats and rats: from sensory acquisition to multiscale maps and navigation

• Large-scale navigational map in a mammal

• Parallel GPU Implementation of Iterative PCA Algorithms

• Human Brain Project

In some sense, one end goal of neuroscience would be to understand the brain enough—both how it is wired and the pattern of neural activity used to produce sensations, store experiences, and create ideas—to allow us to manipulate its activity to allow for entirely new experiences. And while virtual reality is currently very hot in tech—Facebook bought Oculus, Google Glass is still chugging along (and in need of a revamp or new marketing), and Microsoft recently unveiled HoloLens—a real advance would be the ability to directly project scenes and experiences into the human brain, such as a recent study that used electroencephalography (EEG) to record brain signals from one human and use them to induce motor behavior in another human via transcranial magnetic stimulation (TMS) of the motor cortex.(Rao et al., 2014)

On a related note, this would potentially allow us to hijack our knowledge of the brains amazing plasticity to expand our repertoire of senses. Neil Harbisson gave a talk about listening to color, whereby a device allows him to hear different colors, see I listen to color. This uses an eyeborg that converts light into sound. It is easy to imagine adapting such a device to list in for UV, X-ray, Infrared, and other radiation, as is used in a primitive manner with Geiger counters and similar devices. Only, imagine this information could be routed directly to the visual cortex, allowing one to rudimentary see other forms of radiation without blocking visualization of the normal visible light spectrum.

• Microsoft HoloLens

• Google Glass

• Oculus VR - now owned by Facebook

• metaio

• Magic Leap

• A Direct Brain-to-Brain Interface in Humans

• On writing well

• Why academics can’t write

• Why Academics Stink at Writing

• Why Is Academic Writing So Academic?

• In Defense of Academic Writing - a nice counterbalance to the other articles.

The article recommends a couple books for those looking to write clean, concise prose that a reader can easily and logically follow.

• A Rulebook for Arguments

• The Elements of Style

• The New Fowler's Modern English Usage

He also notes several scientific authors whose writing abilities are to be admired and studied: Waltz, Thomas Schelling, James Scott, John Mueller, Deirdre McCloskey, and Charles Krauthammer.

Garret Stuber has a new paper detailing the role of specific lateral hypothalamic (LH) neuron populations in promoting appetitive behavior. Part of the study uses the miniature microscope to image endogenous neural activity in the LH. The resulting analysis is rather simple, but does suggest that consummatory and appetitive behavior induce firing in non-overlapping LH neuron populations.

Their videos have a good deal of brain motion, which will affect proper cell detection and calcium signal extraction techniques; though, their neurons are magnified (owing to the fact that they are using the smaller 0.5mm in diameter lenses instead of the more standard 1.0mm). This is an okay demonstration of the technique, no detailed analysis of the code used by the LH beyond some cells go up and others go down or validation of their cross-session cell alignment is given. Also, the classification of cells into Food-zone excited and Food-zone inhibited cells is not very precise, they only use a ratio between calcium activity in the two zones rather than mutual information or a more precise metric (that takes into account firing rates, time spent in location/how often stimuli are presented, etc.).

A quick note: Figure S6 is rather misleading, as they don’t plot it compared to the animals activity and thus it is unclear whether cells become less responsive to the task or the animal just stops moving (they don’t have a figure plotting a cumulative total movement or distance per time bin as is standard when tracking animal movement in an open field-like environment).

Anyways, as more papers are published using this technique, the rigor in analysis will likely go up as the wow factor decreases and people start looking more at the content of data being produced by the technique.

• Visualizing Hypothalamic Network Dynamics for Appetitive and Consummatory Behaviors

As a nice bonus, the results match what was seen in a paper in the same issue from Kay Tye’s lab.

• Decoding Neural Circuits that Control Compulsive Sucrose Seeking

Choosing a thesis lab, advisor, and scientific problem you want to solve can often be a confusing experience without some guidance and considering that some decisions will end up defining one aspect of your life for several years, it is important to spend the time to get a variety of advice from as many sources as possible. Below are articles going over this decision and what to look out for during rotations and other interactions leading up to deciding on a lab.

• How To Choose a Good Scientific Problem

• Decisions, decisions - great article by Bargmann about the pitfalls to avoid in a scientific career.

• Choosing a Lab

• Choosing The Right Research Adviser

• Choosing a Thesis Lab: Things to consider before and during rotations - great manual from UCSF.

• PhD Students: Should You Switch Labs?

• How to choose a thesis advisor?

• How to succeed in science: a concise guide for young biomedical scientists. Part I: taking the plunge

• How to succeed in science: a concise guide for young biomedical scientists. Part II: making discoveries

• The PhD journey: how to choose a good supervisor

• What I Wish I Knew Before I Entered Grad School

Sometimes it’s good to go back and read the documents you browsed through when first starting graduate school. The main reason is that helps provide perspective and identify where you might be going astray. Below is a useful list of resources along the lines of what it takes to be successful in graduate school.

• 5 Ways to Adjust to Graduate School

• How to Become a Successful Scientist

• No, You're Not an Impostor

• PhD survival guide

• A graduate school survival guide: ``So long, and thanks for the Ph.D!''

• Surviving a bioscience Ph.D.

• Resources for surviving and thriving in graduate school (with a focus on the sciences)

• The importance of stupidity in scientific research

• Vantage Point by Richard Zare: My top 10 principles for success

• What Makes a Great Student in the Lab?

• How to Be a Good Graduate Student -- Succeeding in Graduate School

That is one debate going on in physics as people start to take string theory, multi-verses, and other hypotheses more seriously. The key problem is that some of these cosmological hypotheses are not testable, at least not given what we know about the current rules of physics. This raises a key question, if a theory is sufficiently elegant and mathematically sound, should one need to conduct experiments to prove it? As a biologist, I’m heavily biased towards yes, as that is the foundation of science. However, there are arguments in the other direction as well. See below.

• The Most Dangerous Ideas In Science

• Scientific method: Defend the integrity of physics

• Photoactivatable Neuropeptides for Spatiotemporally Precise Delivery of Opioids in Neural Tissue

• Sebastian Seung's Quest to Map the Human Brain

• Sebastian Seung TED talk

Haruhiko Bito gave a talk today that covered some of his lab’s work using an improved version of RCaMP, a red fluorescent indicator of calcium concentration that is used as a readout of neuronal activity, to do simultaneousness imaging in cortical somatostatin interneurons and excitatory pyramidal cells (using both RCaMP and GCaMP). While the same could be accomplished genetically by crossing Cre versions of those mouse lines with tdTomato reporter lines, this might open up additional experiments that wouldn’t be possible, given that Cre expression during development isn’t always specific, etc.

• Rosetta begins its Comet Tale

• Ambition the film - nice short film advertising the mission.

Gyorgy Buzsaki gave a talk at Stanford today that was fairly enlightening. I’ll just link to several papers he mentioned that are good reads.

• Cognitive neuroscience: Time, space and memory

Might as well post the full list of relevant articles relating to this sub-field that is bound to explode (I’ve already talked about it: see dual photo-stimulation and imaging, freely moving dual photostimulation and imaging, and dual photo-stimulation and imaging, cont'd)

• Two-photon excitation of channelrhodopsin-2 at saturation

• Targeting neurons and photons for optogenetics

• Computer-Generated Holographic Beams for the Investigation of the Molecular and Circuit Function

• Scanless two-photon excitation of channelrhodopsin-2

• Two-photon excitation in scattering media by spatiotemporally shaped beams and their application in optogenetic stimulation

• Functional patterned multiphoton excitation deep inside scattering tissue

• Spatially Selective Holographic Photoactivation and Functional Fluorescence Imaging in Freely Behaving Mice with a Fiberscope

• Simultaneous all-optical manipulation and recording of neural circuit activity with cellular resolution in vivo

• Two-photon optogenetics of dendritic spines and neural circuits

• Simultaneous cellular-resolution optical perturbation and imaging of place cell firing fields

This is an interesting paper that attempts to look at how a 10,000 bit computer would be designed. Keep in mind most computers today are either 32-bit or 64-bit (and in many applications that seems more than enough).

• Computing with 10,000-Bit Words*

Came across this issue of Current Opinion in Neurobiology awhile back, but still find it quite useful.

• Current Opinion in Neurobiology: Theoretical and computational neuroscience

Groups such as Jin Hyung Lee at Stanford and others have looked into using optogenetics and fMRI to probe brain wide neural response to specific perturbations. This could lead to interesting analysis of the effects of stimulating neuromodulatory brain areas—ventral tegmental area, locus coreleus, dorsal raphae, etc.—on whole brain activity. However, always keep in mind that fMRI currently does not measure neural activity, but blood oxygen content, which has been shown in various cases to be weakly correlated, or not at all, to neural activity.

• Optogenetically Induced Behavioral and Functional Network Changes in Primates

Interesting read, looks at re-framing the issue of what the neural code is, e.g. how we make sense of the firing patterns in the brain as they relate to behavior or other cognitive outputs/processes.

• Ecological expected utility and the mythical neural code.

• Light trials for human blindness

• The First International Optogenetic Therapies for Vision Symposium

• 2015 preview: First human trials to cure blindness

• Advancing Genetic Treatments Against Blindness

Also the rest of the NewScientist 2015 science stories are interesting:

• NewScientist: 2015 before it happens

Something to keep in mind now that AI is growing exponentially (see DeepMind) and that if the EU succeed in its initiative to simulate the human brain, the question will arise how do we test that it is ‘intelligent’ in our definition of the term.

• Can Machines Think?

An interesting related question is whether our brains store verbs, nouns, adverbs, and other grammatical constructs differently within the brain, e.g. is the firing patterns in the auditory cortex or Wernicke’s/Broca’s in response to and during the production of speech more distinct between nouns/verbs or solely based on the frequency spectrum produced by those words? Would maybe help answer questions posed long ago about why some word categories are learned sooner than others, e.g. see Why nouns are learned before verbs: Linguistic relativity versus natural partitioning.

From a purely technical standpoint this is an interesting article. That they were able to visualize single laser pulses reflecting off of mirrors and being refracted is super cool. Would be informative to combine this with recent advances in metamaterials that have allowed invisibility to become a reality on the macroscopic scale to look at how light travels through those materials, if possible.

• Single-shot compressed ultrafast photography at one hundred billion frames per second

• Exotic optics: Metamaterial world

At the CNC Annual Symposium last year Hang Lu gave a talk about a collaboration with Kang Shen to help develop high throughput microfluidic devices for screening synaptic deficits in C. elegans. One key take away from the presentation was that machines could detect more subtle changes in synapse morphology, position, and other changes than humans. Further, they could detect changes that induce many weak changes among a series of metrics, something humans find extremely difficult to assess in a rigorous, quantitative manner.

Continuing this theme, at the 2013 Conte Center Neuroscience Symposium, Ulrike Heberlein talked about work she was doing on quantitative descriptions of Drosophila courtship and other behaviors using machine vision in collaboration with the Branson Lab. Along similar lines, another lab at Janelia recently published a high-throughput method of screening for neurons involved in particular Drosophila behaviors (see Discovery of Brainwide Neural-Behavioral Maps via Multiscale Unsupervised Structure Learning). Whether these types of high-throughput characterizations of animal behavior and circuits yield novel or specific insights (beyond just increasing the complexity of the problem) still remains to be seen. Excited to see similar methods applied to research in rodents, which is oftentimes too focused on a single species (Mus musculus or Rattus norvegicus) and a narrow, specific behavior. Curiously, this same complaint about biomedical research has come up more than half a century ago, see Beach’s classic .The snark was a boojum(Beach, 1950)

• Autonomous screening of C. elegans identifies genes implicated in synaptogenesis

• Microfluidics for in vivo imaging of neuronal and behavioral activity in Caenorhabditis elegans

• Microfluidics for the analysis of behavior, nerve regeneration, and neural cell biology in C. elegans

• JAABA: interactive machine learning for automatic annotation of animal behavior

• Discovery of Brainwide Neural-Behavioral Maps via Multiscale Unsupervised Structure Learning

There are many stressors in biology, from experiments that don’t work to charting out the next steps in one’s career. While many of the stressors are common to any job, several are unique to biology because the system being manipulated and studied isn’t the creation of man. And this causes it’s own peculiar problems for our control-obsessed human minds. The below article goes over some common issues biologists encounter while conducting research.

• Stress in Biomedical Research: Six Impossible Things

I remember several years back seeing a presentation on bat 3D place fields at Janelia Farm. You know a presentation is good, or the finding novel enough, that you can remember the room you were in while watching it over three years later. This work comes from Nachum Ulanovsky’s lab, who specializes in studying the coding of space in bats. In the most recent paper they demonstrate that head direction coding of the bat is consistent with a toroidal model of the world, that is if you model the bat’s yaw axis as rotating around the major radius and the pitch as rotating about the minor radius, then this accounts for the patter of place cell activity seen (see figure above or Extended Figure 7a in the text). It also means that the orientation and direction of the bat matters, which would not be the case in a spherical configuration.

• Three-dimensional head-direction coding in the bat brain

• Representation of Three-Dimensional Space in the Hippocampus of Flying Bats - one of the original papers characterizing how three dimensions is represented in bats.

A new paper claims that lifetime risk for getting a particular cancer is highly correlated with the amount of stem cell divisions that a specific tissue undergoes. On the face of it, this makes sense because each time a division occurs, there are chances for mutations to take place that lead to cancer. This has important health implications, as it can imply where treatment should focus, e.g. some cancers might inherently be more treatable through traditional means. This reminds me of an older paper by Piyush Gupta, my old academic advisor at MIT, showing that cancer populations will settle on an equilibrium state. This had implications for treatment, because it suggested that no matter how much you tried to eradicate a particular cancer, it would always come back given the right environment.

• Variation in cancer risk among tissues can be explained by the number of stem cell divisions

• Stochastic state transitions give rise to phenotypic equilibrium in populations of cancer cells.

Was recently listening to Elon Musk talk about his work ethic and principles (see Elon Musk - Work ethics, Principles, Attitude, Failure - Pearls of Advice). In it, he reiterated a point that one notices him mention in many of his interviews and speeches, the idea that taking a first-principles view of the world, as physicists are taught to do, rather than working by analogy can allow you to discover new truths by pointing out logical gaps or counterintuitive thoughts that others missed. He often gives this advice when talking about the formation of SpaceX, that he first reduced the problem of making a space craft to it’s simplest form, that is the raw materials. He then worked up from there, estimating the costs of each component and realizing that rockets in essence where cheap, it is just the rearrangement of the raw materials that is expensive (e.g. like much of manufacturing).

This lead me to start searching for other cases where people attempt to generalize from a physicist’s point of view and apply it to the real world. Below are a collection of articles along those lines.

• Can a biologist fix a radio?—Or, what I learned while studying apoptosis - this is a classic essay on the stages that biomedical research go through and how biologists sometimes approach a problem.

• Why physicists like models, and why biologists should

• A Physicist's Renewed Look at Biology

• More Perfect Than We Imagined: A Physicist's View

• Seeing the Natural World With a Physicist’s Lens

• Conceptual Models and Analytical Tools: The Biology of Physicist Max Delbrück

• Analysis of Transduction Efficiency, Tropism and Axonal Transport of AAV Serotypes 1, 2, 5, 6, 8 and 9 in the Mouse Brain

Was looking into two photon imaging in freely moving animals, it would be useful for multi-colored imaging of different cell-types or other experiments. However, rather than finding a glut of papers, it seems that this fell out of popularity. Several lab members noted that this was partially owing to the fact that instead of using galvos to scan the field of view, most designs vibrated the fiber, which leads to many problems when an animal is moving and is apparently not very reliable. Further, the field of view was never very large, limiting its usefulness in studying large neuronal populations. Anyways, a couple of the papers below:

• Visually evoked activity in cortical cells imaged in freely moving animals

• A Miniature Head-Mounted Two-Photon Microscope: High-Resolution Brain Imaging in Freely Moving Animals

I recently pointed to several papers concerning voltage imaging (see voltage imaging, screening for better probes). A new probe can be added to the list, it is a modified version of ArcLight called Bongwoori. The paper talks a bit about the process they went through to create the probe and go on to characterize it’s properties in hippocampal slices. Seems promising.

• Combinatorial Mutagenesis of the Voltage-Sensing Domain Enables the Optical Resolution of Action Potentials Firing at 60 Hz by a Genetically Encoded Fluorescent Sensor of Membrane Potential

On a slightly different note (building off the voltage sensor news above), there has been a trend toward mesoscopic imaging (see large scale imaging and other papers) and this paper extends that to the visual cortex using voltage indicators. It would be particularly interesting to see someone combine two colors of voltage imaging to look at how membrane fluctuations of glia/astrocytes influence neuronal membrane potentials on a global scale.

• Imaging the Awake Visual Cortex with a Genetically Encoded Voltage Indicator

A paper released last November by Jérôme Lecoq, a postdoc in my thesis lab, uses multiple microscopes to image distal brain regions. While not being able to cover as wide a field of view, he was able to do two-photon imaging and this would allow imaging of interacting brain regions that might be hard to cover in a single field of view.

• Visualizing mammalian brain area interactions by dual-axis two-photon calcium imaging

China is one country where I am still unsure about the status of the research there, e.g. how much of it should be trusted and what the general direction of research there is. For example, while Singapore started off with a fervent passion for basic biomedical research with relatively few top-down restrictions (see Singapore's salad days are over), that began to change. I remember Ian Cheong and others at Temasek Life Sciences while I was there during the summer of 2012 mentioning that there was a shift toward more applied research. Whether that is happening in China and how the vast funds at the government’s disposable are being put to use is still unclear to me. Hopefully I can clear that up soon. In the meantime, this article provide a nice overview of the state and history of molecular and cell biology research. There should be more articles like this on other countries. Especially concerning the pharmaceutical industries in various countries (this probably exists, just need to find an appropriate report).

• In focus: molecular and cell biology research in China

For those wanting to do any work with microcontrollers, the Intel Galileo seems like a good option.

• Intel Galileo

While there are many advantages to imaging—being able to track cells across days, genetic specificity, dual photostimulation and images, and many others—the temporal resolution and several other advantages still make tetrodes a useful tool in any systems neuroscientist’s belt. Below are a couple papers that utilize a technique whereby channelrhodopsin (or another optogentically activatable protein) is expressed in a subset of neurons of interests. Shining light while recording neural activity allows one to identify units that are light responsive, and are thus part of the genetically defined subset that channelrhodopsin was expressed in. Depending on the location and type of cell, there can be asterisks associated with this technique, but it seems to be picking up steam as a viable alternative to calcium imaging, especially as voltage imaging isn’t quite ready for prime time in vivo. Below are some relevant articles in the area.

• Optogenetic identification of striatal projection neuron subtypes during in vivo recordings

• A Guide to In vivo Single-unit Recording from Optogenetically Identified Cortical Inhibitory Interneurons

• In vivo optogenetic identification and manipulation of GABAergic interneuron subtypes

• Targeted optogenetic stimulation and recording of neurons in vivo using cell-type-specific expression of Channelrhodopsin-2

• Amygdala interneuron subtypes control fear learning through disinhibition

While back I linked to several articles that talked about optogenetics and pain (see optogenetics and pain). Wanted to link to several more papers, especially as this particular sub-field has progressed quite a bit since then. In addition, I’ve included several short overviews of the field and a paper that looks at using the technology to restore muscle function after peripheral injury. I’ll talk at length about this area more in the future, but it could help advance the pain field significantly in the coming years while providing hints at possible therapeutic avenues for chronic pain in the coming decades.

• Remote Optogenetic Activation and Sensitization of Pain Pathways in Freely Moving Mice

• Fast-conducting mechanoreceptors contribute to withdrawal behavior in normal and nerve injured rats

• Optogenetic Control of Targeted Peripheral Axons in Freely Moving Animals

• Virally mediated optogenetic excitation and inhibition of pain in freely moving nontransgenic mice

• Optical Control of Muscle Function by Transplantation of Stem Cell–Derived Motor Neurons in Mice

• Optogenetic Recruitment of Dorsal Raphe Serotonergic Neurons Acutely Decreases Mechanosensory Responsivity in Behaving Mice

• A spinal analog of memory reconsolidation enables reversal of hyperalgesia

• Casting light on pain

• Pain: from new perspectives to novel treatments

• Optogenetic Regeneration

• The search for novel analgesics: re-examining spinal cord circuits with new tools

• Nociception and pain: lessons from optogenetics

Was recently watching Deepmind artificial intelligence @ FDOT14, where they demonstrate a new machine learning algorithm that can learn arbitrary rules that govern the movements of pixels on a video screen, also known as learning to play a video game. What is amazing about this is that they don’t give the machine syntactic rules about the games, only the raw video feed and access to a controller, much like a human would have. After an hour or more of training the machine is performing at or above human levels in a variety of Atari games (Pong, Space Invaders, etc.). The white paper can be found at Playing Atari with Deep Reinforcement Learning (also points out the asterisks associated with the algorithm, but still cool nonetheless).

This was partially spurred by recent comments Elon Musk has made regarding the potentially exponential growth in artificial intelligence that could pose a grave danger to humanity in the future should it go unchecked. Nick Bostrom has a rather enlightening book called Superintelligence Paths, Dangers, Strategies that gives an overview of the field of people looking to understand how superintelligence among artificial intelligence might emerge and what we can do about it.

These developments are interesting because it also shines light on a fundamental goal of neuroscience: discover what algorithm the brain uses to allow processing of sensory input to produce an optimal behavioral output. Presumably this would allow us to build better robotic devices, improved search and other algorithms, and develop a host of new technologies. To put things in perspective, a small fruit fly no smaller than a lowercase ‘o’ on the keyboard can navigate complex environments and avoid potential predators using minuscule amount of energy. We currently have nothing machine-wise that can perform at that level and ones that can need orders of magnitude more energy. However, if more and more generalizable algorithms like DeepMind come into play, it may obviate the need to even explore this aspect of neuroscience. This is something people studying reinforcement and other learning paradigms should keep in mind when selling their research, though the issue is still probably a decade or two away from becoming serious enough to cause people to start looking for other jobs.

On a drive down from northern California awhile back Devon Chandler-Brown and I had a conversation about what would happen if a robot committed a crime, say murder or burglary? Who would be responsible? If the owner was responsible, that would open up a whole can of worms, because in essence that would be punishing the parent for the sins of their offspring, something we’ve in general moved past (in the West). However, if the robot was to be punished, how would that play out? If they are not fully self-aware, shutting it down doesn’t send a negative signal to other semi-aware robots (as execution or imprisonment does to humans). On the other hand, if they are fully-aware, they would probably have backed themselves up somewhere, so any punishment is moot since they could just boot up the back-up in case of emergency.

Rather than going into the minutiae of that conversation, I would like to point to several articles that go over this issue in greater detail, as it seems that lawyers are now becoming aware of this fascinating topic and trying to see how it fits within current law.

• Robotics and the Lessons of Cyberlaw

• The Moral Hazards and Legal Conundrums of Our Robot-Filled Future - a great write-up by Greg Miller summarizing a panel discussion at UC Berkeley exploring the issue.

• A Robot Really Committed A Crime: Now What?

• What happens when a software bot goes on a darknet shopping spree?

The popular press is at it again, making bold claims about EP2 research scientists at Stanford have performed. While neurodegeneration is undoubtedly a pressing problem, a healthy skepticism should be made about cures after a recent spat of clinical failures, see the γ-secretase Alzheimer's trials and others. The problem with the γ-secretase trials was that you had to make sure Notch signaling wasn’t radically altered, seeing as it is important signaling cascade. The same might be true for EP2 (see above figure).

• Stanford University researchers found a cure for Alzheimer's disease

• Blocking receptor in brain’s immune cells counters Alzheimer’s in mice, study finds

• Prostaglandin signaling suppresses beneficial microglial function in Alzheimer’s disease models - link to the original article, always best to read the primary source if possible.

• Imaging calcium microdomains within entire astrocyte territories and endfeet with GCaMPs expressed using adeno-associated viruses

• Rapid evaluation of a protein-based voltage probe using a field-induced membrane potential change - demonstrates a method to help screen for voltage probes.

• Imaging neural spiking in brain tissue using FRET-opsin protein voltage sensors - characterizes MacQ, a new sensor.

• All-optical electrophysiology in mammalian neurons using engineered microbial rhodopsins - show all-optical excitation and imaging of cultured neurons with QuasAr and CheRiff.

• Microfluidic Platforms for Single-Cell Analysis

• Microfluidics for single cell analysis

• Single-cell microfluidic impedance cytometry: a review

• The origins and the future of microfluidics

• Microfluidic Device for Single-Cell Analysis

Last month/week I talked about papers from the Tank lab and Emiliani group detailing dual photostimulation and imaging of neural activity (see dual photo-stimulation and imaging and freely moving dual photostimulation and imaging). Adam Packer, from Michael Hausser’s lab, has published a paper in Nature Methods adding to this growing aspect of neuroscience research.

• Simultaneous all-optical manipulation and recording of neural circuit activity with cellular resolution in vivo

Also, the below papers might be of interest as further reading. They demonstrate earlier uses of both patterned excitation profiles with SLMs and laser-scanning excitation.

• Scanless two-photon excitation of channelrhodopsin-2 - demonstrates the use of spatial light modulators for creating patterned excitation profiles, which is used by Adam in his paper.

• Non-redundant odor coding by sister mitral cells revealed by light addressable glomeruli in the mouse - uses laser-scanning to excite individual glomeruli in the olfactory bulb.

Asked Adam about the issues of non-specific photostimulation that both him and the Tank lab are seeing in these experiments, I’ve summarized his response below.

The most relevant figures are figure 3b in the Tank paper and supplementary figures 1 and 5 in Adam’s (see above figure). I had originally thought that the spatial light modulators (SLM) that Adam and co. used in their paper might increase the PSF size leading to reduced specificity in the targeted illumination pattern. However, he mentioned that was not the case and that their PSF size is still determined by objective filling parameters (see The Diffraction Barrier in Optical Microscopy for more). He indicated that it is still unclear how much non-specific stimulation is due direct photostimulation or evoked responses from neurons connected to the targeted neuron.

He did note that photostimulation of dendritic processes was possible (see figure 2 of Packer, 2012) or that out-of-focus stimulation could occur due to channelrhodopsin’s large two-photo excitation profile (see figure 2 of Rickgauer, 2009).

I’ll provide additional updates/thoughts that seem relevant/pertinent as they arise.

The BRAIN initiative provided added fuel to an already growing fire, in this case the trend in neuroscience toward larger collaborations as it has become clear that to tackle the complexity of the brain, expertise from a larger variety of fields needs to be coordinated toward a common purpose. While the benefit of Big Science may be in laying out groundwork for small science to make conceptual advances (as Graur and co. argued was what went wrong with the ENCODE project [along with others], since it tried to do both), there is still the possibility that mixing the two in a single institute could yield a higher probability that the transition would take place. Below are a couple of the new foundations/institutes being pushed along with a slightly older one.

• Paul G. Allen to Give 100 Million to Create Cell Science Institute - Allen already provided a super valuable resources with the Allen Brain Institute, let’s see if it can be repeated with the new cell science institute.

• Simons Collaboration on the Global Brain: Scientific Mission of the Collaboration

• Buck Institute

Several years back the FDA created the Breakthrough Therapies program to help speed the approval process of drugs that will demonstrably provide dramatically better care for a specific group of patients. While the below trial is only a demonstration of using stem cells to treat age-related macular degeneration, it would be interesting to see whether the results from these types of early-stage clinical trials can be fast-tracked into hospitals, should they prove to be safe.

• Pilot safety study of iPSC-based intervention for wet-type AMD

Earlier this year I reported on a talk by Loren Looger that detailed some new technologies coming out of Janelia (see new neuroscience tools from janelia). Below is a short piece talking in more detail about CaMPARI, the .

• Calcium sensor leaves permanent mark in activated neurons

Now that the miniature microscope is in the hands of many labs we should start seeing many of its potential promises: imaging of neural activity in freely moving mice, tracking of neurons across sessions (e.g. days), exploration of genetically or anatomically defined subsets of neurons, and more. A couple of miniature microscope papers have been released recently. One details the effects of an GABA-A agonist (Zolpidem, normally used to treat insomnia) on CA1 neural activity. It would have been more interesting if they also imaged GABAergic neurons in the CA1 to see how they are affected as well. The other paper uses the miniscope to image cancer cells in the vasculature, a quite different goal.

• Zolpidem Reduces Hippocampal Neuronal Activity in Freely Behaving Mice: A Large Scale Calcium Imaging Study with Miniaturized Fluorescence Microscope

• Imaging Circulating Tumor Cells in Freely Moving Awake Small Animals Using a Miniaturized Intravital Microscope

Last month I talked about a paper from the Tank lab detailing dual photostimulation and imaging (DPSI) of neural activity (see dual photo-stimulation and imaging). There is now a new paper from Valentina Emiliani’s group that demonstrates DPSI with freely moving animals. They employ several tricks to get around their non-optimal choice of using ChR2-tdTomato and GCaMP5-G to stimulate and image (they have overlapping excitation spectra). They demonstrate that with some modifications to the imaging, they can reduce background excitation. However, their videos demonstrating dual excitation and imaging are not the most convincing. However, at least they have videos, seems like some imaging papers don’t include any which always makes one a bit suspicious.

• Spatially Selective Holographic Photoactivation and Functional Fluorescence Imaging in Freely Behaving Mice with a Fiberscope

• Wireless Neurosensor for Full-Spectrum Electrophysiology Recordings during Free Behavior - new paper detailing wireless recording in primates.

• A Wireless Multi-Channel Recording System for Freely Behaving Mice and Rats

Some regions of the brain are difficult to access for optical imaging due to their awkward orientation or the location of obstacles (e.g. blood vessels) that make it difficult to access without causing serious damage to overlying tissue or structures. The Tank lab has released a paper outlining a method that uses microprisms to help image the prefrontal cortex (involved in emotion, executive processing, and other higher-level cognitive functions) and medial entorhinal cortex (associated with grid cells that help with navigation of environments). Each structure is either occluded by blood vessels or located at an awkward angle, respectively. This build off previous work, such as a the Levene lab’s paper from last year (see below).

• Chronic Cellular Imaging of Entire Cortical Columns in Awake Mice Using Microprisms - a paper from last year detailing a similar method but instead using it to image different layers of the visual cortex.

• Cellular resolution optical access to brain regions in fissures: Imaging medial prefrontal cortex and grid cells in entorhinal cortex

• The neuronal encoding of information in the brain - rather good review on fundamentals and techniques for analyzing how information is encoded in the brain.

• Neural Representation and the Cortical Code - another great review analyzing different codes used by the cortex.

• Nature promotes read-only sharing by subscribers - this is a nice step forward. There should be a law that dictates for papers that have a certain percentage of government funding (say 50% or greater) they must be offered free to the public.

• Neuroscience and education: myths and messages - a nice primer on some neuroscience myths that pervade in popular culture.

• The warrior in the machine: neuroscience goes to war

• The emergence of functional microcircuits in visual cortex - cool visual cortex paper analyzing how the connectivity rate between L2/3 pyramidal neurons changed after eye opening, specifically between neurons that respond to similar/different stimuli.

• Single-shot compressed ultrafast photography at one hundred billion frames per second

• Top 10 Innovations 2014

There are a myriad of ways to discover research articles: Pubmed (can create personalized rss feeds from searches), Google Scholar (great for finding alternative versions of articles), Web of Knowledge, school library websites, etc. However, there is still a vast trove of papers being published that can be hard to keep up with. Sciencescape hopes to alleviate that problem by providing personalized updates on specific areas of science. Time will tell if this will help people discover and digest research articles better.

• Sciencescape (original, dead)

• Sciencescape (archive link)

• Sciencescape (updated link)

There has been, and continues to be, an explosion in technologies with neuroscience applications. Keeping track of them all can be daunting at times, especially for someone entering the field. I’ve decided to begin assembling a list of neurotechnologies into a living document. Very early draft form is below on Google Doc, still deciding on the best way format and share such a resource.

• Neuroscience Technologies

• Doric Miniaturized Fluorescence Microscopy

Markus Meister has a rebuttal to the recent Science paper that uses a simple assay and projections to predict that humans can discriminate over 1 trillion odors. A brief reading of the paper and rebuttal gives the impression that there is a subtle distinction between how one wants to define ‘discriminate’ in terms of the dimensionality of the sensory systems detectors and how the animal actually perceives the sensations.

For example, subjects report that many complex odor mixtures (composed of  30 individual odor components) actually smell alike.(Weiss et al., 2012) So while a pairwise comparison between many complex mixtures might show discrimination, it does not demonstrate that we actually can perceive all of them differently (in the sense that we could describe the aroma or associate it with a distinct event/object), only that we can tell the difference between mixtures that are nearby in odor-space. This is a subtle point and Markus does a good job teasing apart where the paper appears to have gone off-track and also suggests how one could design an experiment to properly test the question (how many odors can humans detect). Give both papers a read, it’s well worth the effort.

• Humans Can Discriminate More than 1 Trillion Olfactory Stimuli - the original paper.

• Can Humans Really Discriminate 1 Trillion Odors? - the rebuttal.

• Light Field Microscopy for 3D functional neuroimaging

• Simultaneous cellular-resolution optical perturbation and imaging of place cell firing fields

In brief, the paper demonstrate that one can image neural activity (using GCaMP3, an indicator of calcium activity, which is a surrogate for neural activity) and simultaneously stimulate the same neurons individually (using C1V1, a protein that excites cells when red light is shone on them). This idea of measuring neural activity then stimulating those same cells has been around, and several groups have already demonstrated single-cell optogenetic excitation using patterned illumination and other techniques (see below for several examples/reviews).

• Targeting neurons and photons for optogenetics - Hausser and co. give a great overview of some of the concepts tested in the paper.

• Two-photon excitation in scattering media by spatiotemporally shaped beams and their application in optogenetic stimulation

• Computer-Generated Holographic Beams for the Investigation of the Molecular and Circuit Function

• Functional patterned multiphoton excitation deep inside scattering tissue

The paper does a good job showing the technique’s robustness and some limits, but more details on the other optogenetic protein and calcium sensor pairs they claimed to have tested would have been appreciated. The supplemental is rather weak and a video demonstrating the dual stimulating/imaging should have been included (not sure why this isn’t already a requirement for published imaging experiments).

Part of the study looks at using low-power stimulation to perturb network activity and reveal sub-threshold activity of individual cells, which should lead to some interesting applications that build off of Albert Lee's work looking at silent place cells: if combined with recent advances in voltage imaging, this could allow the neuroscience community to address fundamental questions relating to how the intrinsic dynamics of neurons leads to their recruitment during behavior.

This paper, if the technique holds, is going to be a significant new technique in the field and has potential applications beyond neuroscience.¹

• The Craft of Scientific Presentations: Critical Steps to Succeed and Critical Errors to Avoid - goes over the presentation styles of many great scientists/engineers and points out why some succeeded and others failed.

• slide:ology: The Art and Science of Creating Great Presentations

• resonate - Present Visual Stories that Transform Audiences

• Giving science talks - great overview of the general layout of a talk, what to keep in mind, and general tips/pointers.

• Susan McConnell (Stanford): Designing effective scientific presentations (youtube) - was a teaching assistant for Sue and she utilizes many of the concepts she talks about in this video to great effect in her developmental neurobiology class.

• Creating effective slides: Design, Construction, and Use in Science (youtube) - compliments Sue video on presenting.

• Pimp your PowerPoint - just goes over several tips to improve slides.

• Improve your PowerPoint (youtube)

• http://noteandpoint.com/ - pretty nice website for getting inspiration.

• The Presentation Secrets of Steve Jobs

• Effective Presentations in Engineering and Science

• Scientific Presentations The Assertion-Evidence Approach - useful for the videos at the bottom showing students putting presentation approaches into practice.

• We Have Met the Enemy and He Is PowerPoint - an interesting look at where misapplication of a presentation technology can affect how effectively the presentations content is received, discussed, and used.

• Duarte Portfolio - good source of inspiration.

• Wide field-of-view, twin-region two-photon imaging across extended cortical networks

• Visualizing mammalian brain area interactions by dual-axis two-photon calcium imaging

• Thy1-GCaMP6 Transgenic Mice for Neuronal Population Imaging In Vivo

• BRAIN: Launching America’s Next Moonshot

• NIH awards initial \$46 million for BRAIN Initiative research

• Imaging Activity in Neurons and Glia with a Polr2a-Based and Cre-Dependent GCaMP5G-IRES-tdTomato Reporter Mouse - they visualize expression of GCaMP5 under a fluorescence stereoscope, a very low-tech but inexpensive way to check reporter expression patterns.

• Sensory-Related Neural Activity Regulates the Structure of Vascular Networks in the Cerebral Cortex

• Sparseness and Expansion in Sensory Representations

• Multiscale Optical Ca2+ Imaging of Tonal Organization in Mouse Auditory Cortex - cool paper that using epifluoresence imaging to make tonotopic maps in the auditory cortex through a 4mm² field of view. Seems like a useful tool, especially when combined with two-photon imaging to get a more detailed view of neural activity in each region.

• Use of differentiated pluripotent stem cells as replacement therapy for treating disease

Pointing toward physics and saying how they’ve got it figured out is easy, but it doesn’t help in conceptualizing what biologists’ end goals should be. Should we converge on a statistical (thermodynamics), first-principles/observational (classical physics), or some combination of the two? e.g. if i am a big pharma company, would my goals for understanding biology be different than a theoretical systems biologist?

In one case you might just want to do enough experiments that you have large look-up tables for every possible combination of genes and their resulting correlation to cellular action/animal behavior while having secondary databases documenting effects of drugs on these correlations. You won’t need to necessarily know down to the atom exactly how protein A interacts with proteins B-Z, only that protein A has this weak effect and we can correct it (in the case of diseases) with drug A. On the other hand, you might want to build computational models at some level of abstraction—molecules, cells, organ systems, etc.—to simulate the body and thus be able to apply perturbations artificially and see if they match experimental results, validating the underlying models to a degree. These would involve very different technologies and approaches.

Anyways, it seems that Aping Mankind, A Skeptic's Guide to the Mind, and others are good starting points for those wishing to get a picture of the limits of neuroscience and get an alternative view to the hype machine surrounding the USA and Europe’s recent pushes in neuroscience.

This entry is a bit farther afield, but was inspired by a chat with a friend about interstellar travel (partially stemming from a conversation about the excellent the Martian) and how crews would respond if they knew that they would be a stop-gap on a multi-generational mission across the stars. In that sense, it is psychology-related and thus neuroscience related. Bam. Relevant.

So the basic premise is this. Assuming we don’t discover faster-than-light travel and are only ever able to attain several percents the speed of light, it will probably take multi-generational missions to reach other planets. Several questions arise when trying to plan such a mission:

• can people survive the radiation for that long/will offspring be viable?

• what psychological problems will evolve for people who know they will die in the middle of space (and can we simulate/study such conditions on Earth)?

• can you instill the same discipline in second, third, and other generations to prevent politics and other issues from distrupting the mission?

• how do you deal with inbreeding and are all possible matches pre-planned beforehand to reduce this?

• are there fundamental differences in space travel beyond the heliosphere?

• how do you adjust course once inside another star system given that most planets are only detected by the wobble of their host star and their location isn’t exact?

• on more ethical grounds, can we justify terminating or corralling extraterrestrial species if it allows us to properly terraform a planet?

• more to the point, does Manifest Destiny apply to the cosmos (especially when the fate of our species is at stake)? Or should we follow a doctrine similar to the Prime Directive from Star Trek of non-interference?

• what mode(s) of travel are most reasonable and do you need multi-stage systems to assemble the necessary supplies in a cost-efficient manner?

These are just some of the questions that first came to mind and there are many more. It seemed prudent to do a little background research and see what the status was of the field. I’ve compiled a couple books/articles that seem useful and will add to the list as i find more. I might look into contacting those working in this area to get a better idea of the state-of-the-art (especially given Inspiration Mars Foundation and other concrete plans to start testing long-term space flight, ignoring whether these private enterprises will succeed).

• The Starflight Handbook: A Pioneer's Guide to Interstellar Travel

• Interstellar Travel andMulti-Generational Space Ships: Apogee Books Space Series 34

• Deep Space Propulsion: A Roadmap to Interstellar Flight

• Entering Space

• The Case for Mars: The Plan to Settle the Red Planet and Why We Must

• the Martian - this is an excellent survival story and look at near future technologies that NASA might employ during a Mars mission.

• Cosmos: A Personal Voyage - an older TV show, but relevant in providing a layman's outline for our understanding of various aspects of space.

• Islands in the Sky: Bold New Ideas for Colonizing Space

• Spacefaring: The Human Dimension

• Interstellar travel (wikipedia)

• Alone in the Void - good op-ed covering some aspects of the topic, he does have a phrase ``Short of a scientific miracle of the kind that has never occurred[...]'' that makes one want to say ``Challenge accepted.''.

• Time to Plan for a Mission to Alpha Centauri - briefly talks about discovery of an Alpha Centauri B planet and in a wider context, what all the planets discovered by Kepler mean for interstellar travel. He also touches on whether it would be ethical to take-over an already occupied planet.

• Fundamentals of Astrodynamics - a useful reference for those inclined to have a more technical understanding of inter-planetary travel.

• Mining The Sky: Untold Riches From The Asteroids, Comets, And Planets

• Centauri Dreams: Imagining and Planning Interstellar Exploration

Igor Markov has a great review on the fundamental limits to computation. This article reminds me of an older Science article about desalination (The Future of Seawater Desalination: Energy, Technology, and the Environment) that also touches on the fundamental limits in that area by focusing on thermodynamic properties of the process. It is nice that this current paper on computation also talks about the flexibility of what we know regarding limits, thus giving hope for more future advances.

• Limits on fundamental limits to computation

Von Neumann architecture has been the prevailing method for computer design. However, it runs into issues when trying to do massive parallel computing, something the brain is optimized for. A new paper in science explore design of an integrated circuit based on brain architecture. This seems to build off older work at Intel and elsewhere on neuromorphic devices.

• A million spiking-neuron integrated circuit with a scalable communication network and interface

• Proposal For Neuromorphic Hardware Using Spin Devices

Found this Current Opinion in Neurobiology issue from 2012 quite interesting. In particular this figure that shows microcircuit structures and their resulting computations is a nice simplification that should maybe be extended to each new circuit characterized in different brain regions. This has been done extensively for the retina, but doing so in the cortex, striatum, and other regions might aid in a more concrete understanding of those circuits and easier integration of knowledge accumulated from those regions into a larger understanding of how they all work together.

• current opinion in neurobiology: focus on circuits

• microcircuits and their computational roles

• Top Pharma Projects to Watch

The Allen Brain Atlas has been leading the way in collecting, organizing, and sharing systematic data on mouse gene expression and whole brain tracing. Other programs, such as USC’s Mouse Connectome Project, are also helping systematize knowledge in this area. Below are several recent papers that are taking advantage of the explosion in rabies (great overview article) and other retrograde tracers for analyzing regional connectivity and its possible influence on brain function.

• Neural Networks of the Mouse Neocortex

• A Whole-Brain Atlas of Inputs to Serotonergic Neurons of the Dorsal and Median Raphe Nuclei - they also do some electrophysiology to verify the connection between striatal D1 to dorsal raphe and several other structures.

• A mesoscale connectome of the mouse brain

• Organization of Monosynaptic Inputs to the Serotonin and Dopamine Neuromodulatory Systems - this seems to be a follow up to the group's other recent tracking paper: Whole-Brain Mapping of Direct Inputs to Midbrain Dopamine Neurons

• A comprehensive thalamocortical projection map at the mesoscopic level

Below are a couple interesting papers from the last week. Particularly interesting is the last paper looking at the recruitment of different interneuron and motorneuron populations as zebrafish swim faster, as if the organism has a built-in gearbox.

• Single-Cell Phenotyping within Transparent Intact Tissue through Whole-Body Clearing

• Identifying Functional Connections of the Inner Photoreceptors in Drosophila using Tango-Trace

• Network Structure within the Cerebellar Input Layer Enables Lossless Sparse Encoding - interesting paper that explores how encoding properties change as synaptic weights and connectivity are varied

• Presynaptic Partners of Dorsal Raphe Serotonergic and GABAergic Neurons

• A Whole-Brain Atlas of Inputs to Serotonergic Neurons of the Dorsal and Median Raphe Nuclei

• Neurons Are Recruited to a Memory Trace Based on Relative Neuronal Excitability Immediately before Training

• Separate Microcircuit Modules of Distinct V2a Interneurons and Motoneurons Control the Speed of Locomotion

Some recent papers in Nature:

• Genome-scale functional characterization of Drosophila developmental enhancers in vivo - you can almost spot Janelia papers now just by reading the title.

• Cerebellum involvement in cortical sensorimotor circuits for the control of voluntary movements

• `Silent' mitral cells dominate odor responses in the olfactory bulb of awake mice

• A comprehensive thalamocortical projection map at the mesoscopic level

And always nice to see news ways to address neural coding during chronic pain:

• Long-Term Temporal Imprecision of Information Coding in the Anterior Cingulate Cortex of Mice with Peripheral Inflammation or Nerve Injury

In a previous post i talked briefly about quantifying scientific progress. Beyond the literature cited there that explored how resulting scientific productivity compared to funding and review panel scores, there is also another question: how do you quantifying scientific impact? What metrics best account for studies or papers that end up being game changers?

Looking past citations, how do you measure the diffusion of an important paper’s ideas? Especially because they should be useful outside the field in question and thus hard to pinpoint the source when looking through another fields literature. Will we ever have the data to determine where useful ideas pop up? For example, we could have the notes of every scientist and analyze when they read certain articles and when the resulting connecting ideas were formed. And if we did, could we really optimize the discovery process? Those are questions for another day, but for now, the below reading should provide some qualitative and quantitative ideas on how to approach those questions.

One interesting finding from Wang, 2013(Wang et al., 2013) is that for the most highly cited papers, the correlation between the first couple years citations and end citations breaks down. While this makes intuitive sense, the most astounding findings might be ignored because academia is rather conservative in its adoption of new ideas, it does lead one to wonder in what other cases the measures used to review and fund projects start to break down for the most forward thinking ideas, and how one differentiates those from the plain bad projects.

• Quantifying long-term scientific impact

• On the Predictability of Future Impact in Science

• Future impact: Predicting scientific success

• Future Science

• Prediction of highly cited papers

• Anatomy of Scientific Evolution

On another interesting note, the below two papers investigate scientific mobility and impact.

• Driving forces of researchers mobility

• Career on the Move: Geography, Stratification, and Scientific Impact

While the wait is on for useful applications of CLARITY beyond pretty pictures is still on (and the whole endeavor might be superfluous given the new clearing techniques as i've discussed previously), a new paper from the Deisseroth lab seeks to address some of the technique’s problems and provide an overview of the end-to-end needed to get it working (e.g. from clearing to antibody staining to imaging).

Advanced CLARITY for rapid and high-resolution imaging of intact tissues

Other fields, such as clinical medicine, have taken to reporting the effect size, confidence interval, or other statistics that better reflect the biological, and not just statistical, significance of the finding. Below are two good reviews on the usage of effect sizes and other statistical measures other than p-values that should be used when analyzing data.

• It's the Effect Size, Stupid What effect size is and why it is important

• Effect size, confidence interval and statistical significance: a practical guide for biologists

• Cost-Based Metrics for Spike Trains

• Spike train metrics

• Taste Response Variability and Temporal Coding in the Nucleus of the Solitary Tract of the Rat

• Binless strategies for estimation of information from neural data

• From another angle: Differences in cortical coding between fine and coarse discrimination of orientation.

Development and Applications of CRISPR-Cas9 for Genome Engineering

While most studies have a negative control (e.g. just a fluorophore like eGFP/eYFP that doesn’t respond to light + light during behavior), opto-fMRI studies indicate that blue light does cause alterations in brain activity (fMRI response to blue light delivery in the naïve brain: Implications for combined optogenetic fMRI studies). Whether this has more subtle implications that aren’t addressed by behavioral assays used in most studies remains to be seen.

There are many uses for optogenetics where you don’t need to understand how the brain codes for behavior, so long as you obtain a specific behavioral outcome (especially if it will be used in the clinic, similar to the use of deep brain stimulation along with many small molecule and other types of drugs without knowing their exact mechanisms). Much of the hype around optogentics (Why optogenetics deserves the hype) implicitly uses this as a counter to why optogenetics is not overhyped. Fine. Except in scientific research it is being used to elucidate fundamental circuits in behavior (e.g. Dorsal Raphe Neurons Signal Reward through 5-HT and Glutamate) with an implicit/explicit assumption that these circuits encode for or are involved in the computation of specific behaviors.

Given only a specific subset of cells are active for any given stimulus (watch videos with behavior next to neural data from pretty much any two-photon, single-photon, or electrophysiology studies), something commonly attributed to sparse coding, broad spectrum activation is inherently misleading and no study has systematically activated a random subset of brain regions and assayed the mice in a behavioral battery (e.g. Behavioural battery testing: Evaluation and behavioural outcomes in 8 inbred mouse strains) to see how often significant deviations in behavior are seen given certain stimulation parameters (e.g. increase the power, duration, or frequency of the stimulus until an effect is seen).

There are a host of other fundamental problems with optogenetic experiments. I’m merely pointing out one that seems to be mostly ignored in terms of explicitly stating it as a caveat. Whether neural networks respond in a low dimensional ON/OFF fashion as most optogenetic implicitly assume is most likely false based on imaging and other measurements of endogenous neural activity. New tools to more specifically activate a subset of neurons (this is not referring to genetically defined expression, but precise, activity-based subsets) would likely prove more informative with regards to how the brain encodes environmental stimuli, computes, and drives behaviors than the glut of hyper-/hypo-activation optogenetic experiments. More on that in future posts.

Poor Productivity As A Self-inflicted Injury: Who’s Missing The Most Toes, And Why

Who's The Best In Drug Research? 22 Companies Ranked

Pfizer's R\&D Productivity

Munos On Big Companies and Small Ones

R\&D productivity: on the comeback trail

Sheer Economics: How We Got in This Fix

Some of the possibilities of consumer VR are displayed in the following video: Control VR - The Future of VR and Animation

Robust multicellular computing using genetically encoded NOR gates and chemical `wires'
Using temperature to analyse temporal dynamics in the songbird motor pathway
Prediction and validation of the distinct dynamics of transient and sustained ERK activation.
Neuronal generation of the leech swimming movement.
Encoding multiple unnatural amino acids via evolution of a quadruplet-decoding ribosome
Monosynaptic restriction of transsynaptic tracing from single, genetically targeted neurons.

Came across an excellent review (it’s a couple years old) on dendritic computation by London and Hausser over at UCL. Started looking into related papers, below are two interesting papers looking at the role of (different parts of) dendrites in neural computation.

Principles Governing the Operation of Synaptic Inhibition in Dendrites

Evidence for a computational distinction between proximal and distal neuronal inhibition.

chronic, wireless recordings of large-scale brain activity in freely moving rhesus monkeys

In the Schnitzer lab, we regularly record neural activity from freely moving mice during various operant, Pavlovian, fear conditioning, pain, and other behavioral paradigms. Wireless devices are also available for zebra finches and other animals. Now a recent paper has demonstrated recording of neural activity from a freely moving rhesus monkey. Seems they were able to get 5+ years of recordings from some animals, which is pretty amazing, maybe we can start looking at how aging affects gross network activity.

Engineering a memory with LTD and LTP

Many systems neuroscientist will agree that long term depression/potentiation underly some aspects of memory in the nervous system. A new study from the Malinow lab uses optical tools to probe whether fear can be directly manipulated by inducing LTD or LTP. The studies protocol is quite simple and the use of optogenetics over electrical stimulation is unclear (besides optogenetics being in vogue). For example, there have been older studies that look at how inducing LTD/LTP can change learned behavior, such as:

Training-induced and electrically induced potentiation in the neocortex

They also show in the extended data that NMDA receptors are necessary (via MK801 blockade) for the optical conditioned response.

Multipoint-Emitting Optical Fibers for Spatially Addressable In Vivo Optogenetics

One of the main problems with most optical fibers used in optogenetic experiments is that they broadly diffuse light from the tips, potentially exciting nearby brain regions not part of the study. This recent paper from the Sabatini lab over at Harvard provides a next step toward spatially directing light coming out of an optical fiber.

The role of dendrites in auditory coincidence detection
Direct visuomotor transformations for reaching
Ion-channel defects and aberrant excitability in myotonia and periodic paralysis
Axonal delay lines for time measurement in the owl's brainstem
Optimizing Sound Features for Cortical Neurons
Multiplicative computation in a visual neuron sensitive to looming
A neuronal network for computing population vectors in the leech
Flexible control of mutual inhibition: a neural model of two-interval discrimination - really enjoyed this paper and ended up doing a presentation for it.
Influence of dendritic structure on firing pattern in model neocortical neurons
Segregation of object and background motion in the retina
Internally generated cell assembly sequences in the rat hippocampus
Auditory spatial receptive fields created by multiplication
Neuronal correlates of parametric working memory in the prefrontal cortex
Neuronal generation of the leech swimming movement
Evidence for a computational distinction between proximal and distal neuronal inhibition

While it borderlines on the absurd that this wasn’t implemented back in the 80s or 90s, given that governments have been giving citizens unique identifiers for decades, this is a welcome step forward. It would be better if they could go back and add unique identifiers to older authors. This could be automated and mistakes fixed via crowd-sourcing.

boolean logic in bacteria

• Robust multicellular computing using genetically encoded NOR gates and chemical `wires'

learning in culture

• Development, learning and memory in large random networks of cortical neurons: lessons beyond anatomy

• Learning in Networks of Cortical Neurons

http://www.nature.com/nmat/journal/v12/n7/abs/nmat3630.html — A transparent organic transistor structure for bidirectional stimulation and recording of primary neurons

connectivity and weighting

• A synaptic organizing principle for cortical neuronal groups

patterning circuits in vitro

• Constraining the connectivity of neuronal networks cultured on microelectrode arrays with microfluidic techniques: A step towards neuron-based functional chips

• Synaptic plasticity in micropatterned neuronal networks

However, while taking a stats class last year, I happened upon temporal exponential random graphical models (tERGMs). These seemed like a reasonable way of representing the evolving neural activity and getting some idea of the associations between cells during behavior.

Going back to my discussion with Ryan, I’m always open to new ways of quantifying neural activity. In this case, persistent homology seemed like an excellent method assuming we could find some specific way to topologically map the neural patterns into generalizable classes, similar to what was done in the below paper:

Encoding Through Patterns: Regression Tree–Based Neuronal Population Models

And more specifically, it appears that the topological analysis can be performed at different scales, in essence allowing one to determine at what spatial scales the stimuli are being encoded in the neuronal ensemble. Before thinking to deeply about the applications of persistent homology, I asked Ryan to send over a reading list to get up to speed on the subject. See below.

Topological Pattern Recognition for Point Cloud Data
Topology for computing
Three Examples Of Applied and Computational Homology
Simplicial Models and Topological Inference in Biological Systems

Chemogenetic Synaptic Silencing of Neural Circuits Localizes a Hypothalamus->Midbrain Pathway for Feeding Behavior

Always nice to see a paper use DREADDs in an effective manner to dissect a neural circuit, this time the ever fascinating and elusive one involved in feeding behavior. Further, this paper introduces hM4D^nrxn, a modified DREADD that only silences terminals. This allows specific silencing of a regions projection targets while leaving the regions integration and signaling intact. This is useful for circuit tracing, as the effects of having an added G-coupled receptor acting at the soma can have greater unintended consequences compared to one localized at the terminals.

There was a recent paper from Janelia using light-sheet microscopy to image an entire zebrafish at 0.8Hz (see Ahrens, 2013). Another paper from Janelia examined zebrafish embryogenesis using light-sheet microscopy as well (Keller, 2013) while Eric Betzig (also at Janelia) has recently published on methods to image large volumes rapidly (Wang, 2014). Vaziri and Boyden have now published a similar application of light-sheet microscopy, only this time they are imaging both C. elegans and larval zebrafish.

Simultaneous whole-animal 3D imaging of neuronal activity using light-field microscopy

That recent paper reminded me of another paper from last year. Bargmann and Albrecht over at Rockefeller used bright-field microscopy to analyze the neuronal activity of around 20 C. elegans simultaneously (Larsch, 2013). My lab is pursuing a similar project, only this one will image the brains of flies (e.g. D. melanogaster). A similar project was published that used genetically driven expression of channelrhodopsin in Drosophila to help tease apart neural correlates of behavior (see Vogelstein, 2014). Whether large-scale projects like these will yield biological conclusions remains to be seen. However, it seems like they should follow the big-science as a map, e.g. like the Human Genome project was.

Application of Tissue Clearing and Light Sheet Fluorescence Microscopy to Assess Optic Nerve Regeneration in Unsectioned Tissues

Lessons learned from the fate of AstraZeneca's drug pipeline: a five-dimensional framework

SB Laughlin. A simple coding procedure enhances a neuron's information capacity. Z Naturforsch, 36c:910-912, 1981.

Lettvin, J.Y., et al.. What the Frog's Eye Tells the Frog's Brain, Proc. Inst. Radio Engr. 47:1940-1951, 1959.

Ballard, DH, Cortical connections and parallel processing: Structure and function, in Vision, in Brain and cooperative computation, pp 563-621, 1987, Arbib, MA and Hanson AR (Eds).

Y Weiss, et al.. Motion illusions as optimal percepts. Nature Neuroscience, 5(6):598 604, 2002.

BA Olshausen and DJ Field. Sparse coding of sensory inputs. Current Opinion in Neurobiology, 14:481 487, 2004.

T Poggio, V Torre, and C Koch. Computational vision and regularization theory. Nature, 317:314 319, 1985.

Girosi, Federico, Michael Jones, and Tomaso Poggio. Regularization theory and neural networks architectures. Neural computation 7.2 (1995): 219-269.

AA Stocker and EP Simoncelli. Noise characteristics and prior expectations in human visual speed perception. Nature Neuroscience, 9(4):578 585, 2006.

Marr, D., and T. Poggio. Cooperative Computation of Stereo Disparity. Science, 194, 283-287, 1976.

Rumelhart, D. E., et al.. Learning representations by back-propagating errors. Nature, 323, 533–536.

Hinton, Geoffrey E., and Steven J. Nowlan. How learning can guide evolution. Complex systems 1.3 (1987): 495-502.

Hinton, G. E. and Plaut, D. C. Using fast weights to deblur old memories. Proceedings of the Ninth Annual Conference of the Cognitive Science Society, Seattle, WA

Becker, S. and Hinton, G. E. A self-organizing neural network that discovers surfaces in random-dot stereograms. Nature, 355:6356, 161-163

Ackley, David H., Geoffrey E. Hinton, and Terrence J. Sejnowski. A learning algorithm for Boltzmann machines. Cognitive science 9.1 (1985): 147-169.

Durbin, Richard, and David Willshaw. An analogue approach to the travelling salesman problem using an elastic net method. Nature 326.6114 (1987): 689-691.

Douglas, Rodney J., Kevan AC Martin, and David Whitteridge. A canonical microcircuit for neocortex. Neural computation 1.4 (1989): 480-488.

Swindale, N. V. A model for the formation of orientation columns. Proceedings of the Royal Society of London. Series B. Biological Sciences 215.1199 (1982): 211-230.

Zohary, E, Shadlen, MN and Newsome, WT (1994). Correlated neuronal discharge rate and its implications for psychophysical performance. Nature 370:140-143.

Hopfield, John J. Neural networks and physical systems with emergent collective computational abilities. Proceedings of the national academy of sciences 79.8 (1982): 2554-2558.

Hopfield, John J. Neurons with graded response have collective computational properties like those of two-state neurons. Proceedings of the national academy of sciences 81.10 (1984): 3088-3092.

Hodgkin, Alan L., and Andrew F. Huxley. A quantitative description of membrane current and its application to conduction and excitation in nerve. The Journal of physiology 117.4 (1952): 500. - And might as well read the rest of their 1952 series.

Song, Sen, and Larry F. Abbott. Cortical development and remapping through spike timing-dependent plasticity. Neuron 32.2 (2001): 339-350.

Buonomano, Dean V., and Michael M. Merzenich. Temporal information transformed into a spatial code by a neural network with realistic properties. Science. 1995 Feb 17;267(5200):1028-30.

Wilson, Hugh R., and Jack D. Cowan. Excitatory and inhibitory interactions in localized populations of model neurons. Biophysical journal 12.1 (1972): 1-24.

Pouget A, Deneve S, Duhamel JR (2002) A computational perspective on the neural basis of multisensory spatial representations. Nat Rev Neurosci. 3: 741-747.

Olshausen BA, Field DJ. 1996. Emergence of simple-cell receptive field properties by learning a sparse code for natural images. Nature 381:607-9.

Atick, Joseph J. Could information theory provide an ecological theory of sensory processing?. Network: Computation in neural systems 3.2 (1992): 213-251.

Milo, Ron, et al. Network motifs: simple building blocks of complex networks. Science 298.5594 (2002): 824-827.

Song, Sen, et al. Highly nonrandom features of synaptic connectivity in local cortical circuits. PLoS biology 3.3 (2005): e68.

Chklovskii, Dmitri B., B. W. Mel, and K. Svoboda. Cortical rewiring and information storage. Nature 431.7010 (2004): 782-788.

Traub, Roger D., and Richard Miles. Neuronal networks of the hippocampus. Vol. 777. Cambridge University Press, 1991.

Whittington, Miles A., Roger D. Traub, and John GR Jefferys. Synchronized oscillations in interneuron networks driven by metabotropic glutamate receptor activation. Nature 373.6515 (1995): 612-615.

Pinsky, Paul F., and John Rinzel. Intrinsic and network rhythmogenesis in a reduced Traub model for CA3 neurons. Journal of computational neuroscience 1.1-2 (1994): 39-60.

De Schutter, Erik, and James M. Bower. An active membrane model of the cerebellar Purkinje cell. I. Simulation of current clamps in slice. Journal of neurophysiology 71.1 (1994): 375-400.

Markram, Henry. The blue brain project. Nature Reviews Neuroscience 7.2 (2006): 153-160.

Markram, Henry, Yun Wang, and Misha Tsodyks. Differential signaling via the same axon of neocortical pyramidal neurons. Proceedings of the National Academy of Sciences 95.9 (1998): 5323-5328.

Treves, Alessandro, and Edmund T. Rolls. Computational analysis of the role of the hippocampus in memory. Hippocampus 4.3 (1994): 374-391.

Rolls, Edmund T., Alessandro Treves, and Edmund T. Rolls. Neural networks and brain function. (1998).

London, Michael, and Michael Häusser. Dendritic computation. Annu. Rev. Neurosci. 28 (2005): 503-532.

De La Rocha, Jaime, et al. Correlation between neural spike trains increases with firing rate. Nature 448.7155 (2007): 802-806.

Steriade, Mircea, David A. McCormick, and Terrence J. Sejnowski. Thalamocortical oscillations in the sleeping and aroused brain. Science 262.5134 (1993): 679-685.

Bell, Anthony J., and Terrence J. Sejnowski. The `independent components' of natural scenes are edge filters. Vision research 37.23 (1997): 3327-3338.

Churchland, Patricia Smith, and Terrence J. Sejnowski. The computational brain. The MIT press, 1992.

Bhalla, Upinder S., and Ravi Iyengar. Emergent properties of networks of biological signaling pathways. Science 283.5400 (1999): 381-387.

Zador, Anthony, Christof Koch, and Thomas H. Brown. Biophysical model of a Hebbian synapse. Proceedings of the National Academy of Sciences 87.17 (1990): 6718-6722.

Brown, Thomas H., Edward W. Kairiss, and Claude L. Keenan. Hebbian synapses: biophysical mechanisms and algorithms. Annual review of neuroscience 13.1 (1990): 475-511.

Destexhe, Alain, Zachary F. Mainen, and Terrence J. Sejnowski. Synthesis of models for excitable membranes, synaptic transmission and neuromodulation using a common kinetic formalism. Journal of computational neuroscience 1.3 (1994): 195-230.

Liley, A. W., and K. A. K. North. An electrical investigation of effects of repetitive stimulation on mammalian neuromuscular junction. Journal of Neurophysiology 16.5 (1953): 509-527.

Betz, W. J. Depression of transmitter release at the neuromuscular junction of the frog. The Journal of physiology 206.3 (1970): 629.

Abbott, L. F., et al. Synaptic depression and cortical gain control. Science 275.5297 (1997): 221-224.

Tsodyks, Misha V., and Henry Markram. The neural code between neocortical pyramidal neurons depends on neurotransmitter release probability. Proceedings of the National Academy of Sciences 94.2 (1997): 719-723.

Hopfield, John J., and A. V. Herz. Rapid local synchronization of action potentials: Toward computation with coupled integrate-and-fire neurons. Proceedings of the National Academy of Sciences 92.15 (1995): 6655-6662.

Von Der Malsburg, Christoph. The correlation theory of brain function. Springer New York, 1994.

Willshaw, David J., and Christoph Von Der Malsburg. A marker induction mechanism for the establishment of ordered neural mappings: its application to the retinotectal problem. Philosophical Transactions of the Royal Society of London. B, Biological Sciences 287.1021 (1979): 203-243.

Srinivasan, Mandyam V., Simon B. Laughlin, and Andreas Dubs. Predictive coding: a fresh view of inhibition in the retina. Proceedings of the Royal Society of London. Series B. Biological Sciences 216.1205 (1982): 427-459.

Loren Looger, group leader at Janelia, gave a rather interesting talk today at Stanford. He went over several new technologies being developed at Janelia and the talk had an interesting, choose-your-own adventure style, a nice twist on the normal talk style. In addition, he had executive summaries for parts of the talk that he could not go over in detail. Overall, it was an excellent talk and from the discussion afterwards, it is clear that he understands and is willing to discuss the limitations of the various technologies he’s developing and the ones currently in use.

calcium sensors

Nothing crazy new here. He mentioned that GCaMP7 is in the works and will have a lower baseline background (F0) and better SNR.

neurotransmitter sensors

iGluSnFR (intensity-based glutamate-sensing fluorescent reporter, see Marvin, 2013) can be used as a measure of glutamate release. The kinetics of their most recent version are extremely fast, such that a slow confocal or two-photon might not show any activity because the frame-rate is too slow or averaging is washing the signal out. He mentioned that Lin Tian is working on a split trans-synaptic version, which should allow measurement of synaptic strength between regions, which could be a real boon for people studying plasticity in various paradigms.

The sensor has been used in two recently published papers:
Kainate Receptors Mediate Signaling in Both Transient and Sustained OFF Bipolar Cell Pathways in Mouse Retina
Two-Photon Imaging of Nonlinear Glutamate Release Dynamics at Bipolar Cell Synapses in the Mouse Retina

In addition to a glutamate sensor, he mentioned that they are attempting to develop dopamine, acetylcholine, adenosine, and other neurotransmitter receptors to help people answer questions pretaining to those systems, these new sensors should prove useful to Drosophila researchers where the main neurotransmitter is acetylcholine.(O’Kane, 2011) The dopamine sensor reminds me of Alan Jasanoff’s recent dopamine contrast agent paper, Lee, 2014, which uses a sensor based on directed evolution in Shapiro, 2010. While fMRI has several advantages, it is still limited by needing the animal to remain motionless, which is where the sensors could prove useful in combination with the miniature microscopes in Mark Schnitzer’s lab or other imaging technologies in the pipeline.

experiments A brief idea: the DA sensor could prove extremely helpful in deciphering the role of dopamine in both limbic and cortical (especially prefrontal cortex) structures as it relates to modulation of neuronal activity during learning, reward, addiction, and pain. While the cellular effects of the various dopamine receptors are known to a degree, precisely how local dopamine effects neuronal activity has been unavailable. If the dopamine sensor ends up being green, one could take advantage of RCaMP (see Akerboom, 2013) to simultaneous image local dopamine release and neuronal activity.

dual fluorescence and EM imaging

EosFP is an osmium resistant fluorophore that should be useful for helping do tracing and then using scanning electron microscope (SEM) or transmission electron microscopy (TEM) to do more detailed tracing. This can allow for error correction because the sparse labeling via the fluorescent protein can be used to see if there has been a register shift or some other problem. They are trying to make them resistant to EPON, a particular epoxy used to help keep sample rigid during slicing.

imaging fine processes

smFP (spaghetti monster fluorescent proteins, yes he has a sense of humor when naming his technologies) can be used to detect finer details when doing tracing or other studies. They basically have added FLAG, HA, myc, and potentially other protein tags that have been used for decades to allow pull-down in biochemical assays, among a variety of other uses. The tech doesn’t seem revolutionary, but appears to be a useful tool for dissecting neural circuits or things going on in areas with finer synapses and detail. For example, showed CA3 spine thorns, which are much more complex than normal dendritic spines. The data looked quite convincing.

calcium integrators

CaMPARI (calcium modulated photoactivatable ratiometric integrator, also named from the liqueur Campari) is the name given to the new sensor. It works by taking advantage of the fact that some sensors photoconvert to a new state upon being hit by photons while others (like GCaMP calcium sensors) change state when calcium concentration changes.

problems Looger noted several drawbacks of the technique. Because of calcium dynamics, an experimenter can’t use CaMPARI to measure precise spike timing, as spikes before turning on the UV light can lead to calcium transients that leak into the light-on time. In addition, in order to photoswitch, UV light needs to be used, with two obvious problems: this can be damaging to the tissue being manipulated and is amenable to deep brain without use of a light probe or another manipulation.

applications He showed it worked with linear response by stimulating rat hippocampal neurons. Futher, using a flight-response paradigm in zebra fish (e.g. poke them with a pencil to initiate a flight response), they showed that they could get activation in freely moving fish but not those that were immobilized. Further, they partnered with Karel Svoboda to demonstrate the tech could detect direction selectivity of cells in the mouse visual cortex. Lastly, he noted that pilot experiments implied that they could visualize labeled line of the Drosophila olfactory circuit, e.g. olfactory receptor neuron to projection neuron and onward to tertiary and quaternary neurons. Could be quite interesting. Also, he mentioned that it might be possible to use three-photo imaging (this new report gives a nice overview) to do non-invasive, deep imaging of activated neurons in more superficial structures.

experiments I could easily envision several interesting experiments, for example one could look at either cocaine administration or morphine sensitization and for each dose, apply UV light. Then sacrifice the animal and FACS sort the cells then sequence to see whether there are specific cell types that are activated during drug administration (or take advantage of transgenetic Cre animals and td-Tomato reporter lines to just do two-photon alignment). Pain is a field that seems a fair bit behind in terms of systems neuroscience analysis of supraspinal regions. CaMPARI would allow an exploratory look at regions activated during noxious stimuli, pain relief or chronic pain. This might be several order of magnitude more sensitive than TRAP or similar technologies based on immediate early genes. Though, this tech is at the moment more invasive.

designed genetic switches

Apparently DuPont called Looger a decade or so ago to help them develop a system whereby they could turn on specific genes within crops. The incentive was that they had many transgenetic crops or variants that had drought, pest, etc. resistance but they could not establish a line because the F0 was sterile. They initially were looking at the commonly used inducible gene system, tetracycline. However, for obvious health reasons, this could not be put into widespread agriculture use. Thus, they started looking at several pesticides, one of which was chlorsulfuron, which happened to chemically look the nearest to tetracycline. Looger was able to show that they could use guided computational design and directed evolution to make tetracycline repressor become specific to chlorsulfuron (see the patent US 8257956 B2 - Sulfonylurea-responsive repressor proteins). They can induce expression of a reporter protein either use a spray or root absorption.

GWAS 2.0

Looger seemed to focus on the fact that many GWAS studies point toward a general region in the genome or SNPs that are correlated with a particular behavior, but that they just leave it at that, which isn’t very satisfactory from a mechanistic standpoint. He wants to start seeing if there is a systematic way to look at where exactly the SNPs are occurring, enhancers, binding sites, etc. and what the biochemical consequences of these are. Then systematically seeing the changes that occur. He has already started along this track with a recent paper:

Allelic heterogeneity in NCF2 associated with systemic lupus erythematosus (SLE) susceptibility across four ethnic populations

Two recent papers in Science and one in Nature Medicine build on previous work (see Villeda, 2011) looking into the effects of young blood on aging mice. One, from Amy Wagers’s group at Harvard, identifies GDF11 as a possible molecular component of young blood crucial for the anti-aging effects seen, specifically focusing on skeletal muscles. The other, coming out of Lee Rubin’s lab in collaboration with the Wagers lab, also illustrate the role of GDF11, but this time looking at its effect on vasculature and neurogenesis. Lastly, work from the Wyss-Coray lab at Stanford focus on using microarray analysis to identify genes altered by parabiosis of young and old animals. They identify Egr1, an immediate early gene associated with neuronal activity, and Creb as displaying increased expression and phosphorylation status, respectively.