This separation means that computation is often rate-limited by the speed at which information can be transferred between memory and the CPU, referred to in John Backus’s Turing Award lecture as “the von Neumann bottleneck”. In Danny Hillis’s book “The Connection Machine” (which I highly recommend), he argues that the von Neumann architecture additionally unnecessarily slows down computation because most of the silicon in a computer is sitting there unused most of the time when it’s being used to store memories which are not currently being accessed (Hillis proposed solution was massively parallel memory-oriented computing).

In addition to Hillis’s argument that the von Neumann design is temporally inefficient, this NYTimes commentary on Ranganathan’s article argues that it is energy inefficient, citing a panel that found that the energy cost of moving data between memory and processors is more than 10x the energy cost of the processing itself (and possibly more than 100x). In other words, massively parallel memory-oriented computing, which seems to be how the brain works, may be both faster and more energy-efficient that von Neumann computing (what’s the catch? You have to write massively parallelizable algorithms to run on it). The energy-efficiency part of this isn’t too surprising, as evolution had a lot of selection pressure to optimize for low energy use. It’s neat though.

Parthasarathy Ranganathan. From Microprocessors to Nanostores: Rethinking Data-Centric Systems. IEEE Computer January 2011, p. 39-48.

NYTimes summary: Remapping Computer Circuitry to Avert Impending Bottlenecks – NYTimes.com.

This paper provide a model of the slime mold’s network construction algorithm.

“When [Physarum] grows on a nutrient-rich substratum, it covers the surface as a coherent layer (like a pancake). If nutrition becomes limited, it forms fenestrae and finally transforms into a network of interconnected veins that enclose the entire cytoplasmic volume (see the figure). Each vein is a gel-like tube covered by a cell membrane and contains a core of fluid cytoplasm. By rhythmic contraction of its cytoskeleton, cytoplasm is continually pumped through these veins, and this continuous mixing seems to be the reason why all nuclei proceed synchronously through the cell division cycle. The network architecture is highly dynamic. Veins change in thickness, they may form and vanish again, and the plasmodium as a whole can crawl over its substratum, moving over centimeters in a couple of hours. Plasmodia usually do not dissociate. If food sources are spatially separated, such as oat flakes scattered over a wet surface, the plasmodial veins attempt to connect these food sources along the shortest possible pathways, even finding optimized paths through a maze (2). How this optimization is performed in terms of molecular mechanisms remains a challenging question.”

[The model is] “based on feedback loops between the thickness of each tube and internal protoplasmic flow (18–22) in which high rates of streaming stimulate an increase in tube diameter, whereas tubes tend to decline at low flow rates (23). The initial shape of a plasmodium is represented by a randomly meshed lattice with a relatively fine spacing … The edges represent plasmodial tubes in which protoplasm flows, and nodes are junctions between tubes.”

(commentary)

Atsushi Tero, Seiji Takagi, Tetsu Saigusa, Kentaro Ito, Dan P. Bebber, Mark D. Fricker, Kenji Yumiki, Ryo Kobayashi, and Toshiyuki Nakagaki. Rules for Biologically Inspired Adaptive Network Design. Science 327 (5964), 439. (22 January 2010)

Drazen Prelec. A Bayesian Truth Serum for Subjective Data. Science 15 October 2004: Vol. 306. no. 5695, pp. 462 – 466. DOI: 10.1126/science.1102081

Here’s an example that demonstrates the crux of the method. The question is, “Is Picasso your favorite visual artist?” — assume that Picasso lovers are a minority. We want to develop an incentive system that gives Picasso lovers an incentive to answer truthfully that Picasso is their favorite, even though they are in the minority:

People who, for example, rate Picasso as their favorite should — and
usually do … — give higher estimates of the percentage of the population who shares
that opinion, because their own feelings are an informative `sample of one’ …. It
follows, then, that Picasso lovers — who have reason to believe that their best estimate of
Picasso popularity is high compared to others’ estimates — should conclude that the true
popularity of Picasso is underestimated by the population. Hence, one’s true opinion is
also the opinion that has the best chance of being surprisingly common.

Based off this idea, the method rewards people for giving “surprisingly common” answers. Each person is asked not only for their own answer, but also to predict the frequency of each answer in the population. The following equation is used to calculate the reward to give each person (equation 2, “score for respondent r” in the paper, page 5 of the PDF linked above):

log ((the actual frequency of this guy’s answer in the poll)/(the geometric mean of the predicted frequency of this guy’s answer in the poll))
+
alpha * sum over all answers (the actual frequency of this answer in the poll * log((this guy’s prediction of the frequency of this answer in the poll)/(the actual frequency of this answer in the poll)))

where alpha is a parameter between 0 and 1.

The first term in the reward rewards people for giving “surprisingly common” answers. The second term rewards people for giving accurate predictions of the frequency of answers.

The paper goes on to show that, given this reward function, truth-telling is a Nash equilibrium, and furthermore that for sufficiently small alpha, this equilibrium Pareto-dominates expected scores in other equilibria. It also discusses things that can go wrong, and what to do about them.

A few days later, Henry Markram, leader of the Blue Brain Project at EPFL, sent off an e-mail to IBM CTO Bernard Meyerson harshly criticizing the IBM press release, and cc’ed several reporters. This brought a spate of shock media into the usually placid arena of computational neuroscience reporting, with headlines such as “IBM’s cat-brain sim a ‘scam,’ says Swiss boffin: Neuroscientist hairs on end”, and “Meow! IBM cat brain simulation dissed as ‘hoax’ by rival scientist”.  One reporter chose to highlight the rivalry as cat versus rat, using the different animal model choice of the two researchers as a theme.  Since then, additional criticisms from Markram have appeared online.

Find out more after the jump.

In the aftermath, IBM has stood behind the announcement, citing for Network World their team’s involvement with “Stanford University, University of Wisconsin-Madison, Cornell University, Columbia University Medical Center, University of California-Merced and Lawrence Berkeley National Laboratory” as defense.  Who are the researchers they are standing behind?  According to Modha’s blog, they are:

• Stanford University: Brian A. Wandell (Prof of Psychology, Electrical Engineering), H.-S. Philip Wong (Prof of Electrical Engineering)
• Cornell University: Rajit Manohar (Prof of Electrical Engineering)
• Columbia University Medical Center: Stefano Fusi (Prof of Theoretical Neuroscience)
• University of Wisconsin-Madison: Giulio Tononi (Prof of Psychiatry)
• University of California-Merced: Christopher Kello (Prof of Cognitive Science)

For this neurodude, it is interesting how this disagreement may be symbolic of the gap that still remains between neuroscience and AI.  Markram is a neuroscientist turned technologist, while Modha is a computer engineer who wants to derive technological insight from biological  systems.  They are approaching the ideal of reverse engineering the brain from very different perspectives, and its only natural that they value different milestones.  The IBM team, even with the additional professors on their team, still lacks mainstream neuroscientists to help validate their claims.  That being said, the public realization of this could be a positive thing for both fields.  Although some frustration has resulted from this, this could be a great opportunity for the breakdown of walls between these fields.

In the end though, it does seem like Markram has a point.  The IBM press release clearly went too far.  Whether the angry public e-mail was the best strategic way to make the point remains to be seen.  It will be interesting to see what the next move from the IBM team will look like.

I’m a fan of it because it is an open-access journal featuring a “tiered system” and more.  From their website:

The Frontiers Journal Series is not just another journal. It is a new approach to scientific publishing. As service to scientists, it is driven by researchers for researchers but it also serves the interests of the general public. Frontiers disseminates research in a tiered system that begins with original articles submitted to Specialty Journals. It evaluates research truly democratically and objectively based on the reading activity of the scientific communities and the public. And it drives the most outstanding and relevant research up to the next tier journals, the Field Journals.

I’m a big fan of the variety of specialty journals they have:

• Aging Neuroscience
• Behavioral Neuroscience
• Cellular Neuroscience
• Computational Neuroscience
• Enteric Neuroscience
• Evolutionary Neuroscience
• Human Neuroscience
• Integrative Neuroscience
• Molecular Neuroscience
• Neural Circuits
• Neuroanatomy
• Neuroenergetics
• Neuroengineering
• Neurogenesis
• Neurogenomics
• Neuroinformatics
• Neuromethods
• Neuropharamacology
• Neuroprosthetics
• Neurorobotics
• Synaptic Neuroscience
• Systems Neuroscience

Just last night, I watched the tense but amazing film The Hurt Locker (don’t let the name disuade you, see the phenomenal Metacritic rating), which is about a bomb disposal squad during the recent Iraq War. There is one particularly stirring scene with a suicide bomber who claims that he was forced to wear a vest with explosives and doesn’t want to go through with it. The difficulty in the limited time before the bomb explosion revolves around whether to actually trust the man and the challenge of trusting someone when neither party speaks the other’s language. You can certainly at least understand (putting aside the ethics of war itself) why governments are interested in detecting nonverbal trust cues.

Details about the IARPA call for proposals are after the jump.

IARPA is soliciting submissions on the following areas aimed at
addressing the challenges of defining, understanding, and ultimately
detecting valid, reliable signatures of trust in humans:

1.) Different kinds of trust and what, if any, kinds of
neurophysiological signals might be associated with them. IARPA seeks
to understand the different manifestations that trust may take (i.e.
swift trust, conditional trust, unconditional trust, etc.) as well as
the different neurophysiological processes associated with one or more
of these kinds of trust.

2.) New models of neural systems and patterns of neural activation
related to different kinds of trust and associated neurophysiological
signatures of those activation patterns. IARPA seeks to understand the
degree to which the neural-bases of trust(s) may assist in detecting
peripheral signals of trust and trustworthiness under different
conditions.

3.) Potentially novel preconscious signals or combinations of signals
- neural, endocrine, physiological, behavioral, etc. – that may be
indicative of trust or trustworthiness among people in different
contexts. IARPA seeks to elucidate signals and neurobiological
processes that humans may use for assessing trust, but which are not
yet – or are currently poorly – understood.

4.) New sensor technologies or combinations of technologies that can
assist in detecting subtle but valid and reliable changes in
neurophysiological states that may be indicative of trust among
humans. IARPA seeks to explore the feasibility of using technology to
amplify systems that humans have evolved to preconsciously assess
trust in others.

5.) Novel, ecologically-valid, but ethical “trust-based” protocols
designed to assess the validity and reliability of potential trust
signals among two or more humans. IARPA seeks to develop new, more
sophisticated processes for understanding near real-time human
preconscious assessment of trust in near real-world circumstances.

Atul Gawande also recently wrote a New Yorker article about treating phantom itch with Ramachandran’s mirror box. I found this part of Gawande’s article on statistical inference in perception most interesting:

You can get a sense of this from brain-anatomy studies. If visual sensations were primarily received rather than constructed by the brain, you’d expect that most of the fibres going to the brain’s primary visual cortex would come from the retina. Instead, scientists have found that only twenty per cent do; eighty per cent come downward from regions of the brain governing functions like memory. Richard Gregory, a prominent British neuropsychologist, estimates that visual perception is more than ninety per cent memory and less than ten per cent sensory nerve signals. When Oaklander theorized that M.’s itch was endogenous, rather than generated by peripheral nerve signals, she was onto something important.

I’m not familiar with this field but I wonder if anyone has tried to quantify what percent of our conscious experience that we normally believe to be 100% due to sensory input is actually recall from memory/inference based on past observation. Also, can this percentage adaptively change? Perhaps there are situations where the brain chooses to rely more heavily on memory and other cases where it relies more on primary sensory input.

Interviews Boyden and Deisseroth. Follow the link a video of an optogenetically controlled mouse.

I think this chapter might be helpful for neuroscientists outside of the lab. Often a dinner table discussion has moved to the idea of “quantum consciousness” or “quantum free will”. Often, someone will mention Roger Penrose, who has become something of a poster boy for this idea that quantum indeterminacy (eg. Heisenberg’s uncertainty principle) is one possible way that free will is really free. And then, people look around and say, “Well, you’re a neuroscientist. Do we have free will?” (And that’s when I take another big drink or bite while I figure out something semi-coherent to say.)

Sompolinsky does a nice job of evaluating such claims (in the end, he says we cannot rule out the possibility that the brain is an indeterministic system but it seems unlikely) and provides nice scientific insight. In his view, it is far more likely that the brain’s apparent randomness (eg. individual cell spike rasters vary across repeated presentations of the same stimulus) is more simply explained by thermal noise (think of varying channel gating properties) and chaotic brain dynamics. (Recall, a chaotic system is still deterministic; it simply exhibits aperiodic behavior due to exquisite sensitivity to initial conditions. It is difficult to predict the long-term behavior of chaotic systems. The more we know the initial conditions in detail, the better our prediction.) On the other hand, he argues that the relevant length and time scales for neurons (micrometers and milliseconds) are far larger by many orders of magnitude than those of quantum noise. Chaos might amplify such quantum events, but this is far from being the simplest, most parsimonious explanation. Given the current level of neuroscience understanding, this is almost idle speculation. Regardless of the (in)determinacy of the world, Sompolinsky effectively argues against any non-physical, purely mental (ie. dualistic) agent of causation.

Thus, in sum, the world and our brains might not be determined but, even given that, there’s no reason to believe we have any causative ability to change things in the sense of traditional free will. These observations seem right on the mark to me. I hope they bring some insight for others. Or at least a way to fend off the dinner-table-free-will-conversation barrage of questions.