Cognitive Enhancement Debate

Zack Lynch posted a video debate about cognitive enhancement on his blog. I just wanted to mention a few of my thoughts on this topic. I think currently a lot of the so called memory improving drugs leave a lot to be desired. I would say that the average person probably would not want to take a drug which caused increased activation of acetylcholine receptors. Dopamine reputake inhibitors can improve concentration and certain aspects of cognition, but they are also not ideal drugs in any way. They can be addictive and may also increase other unwanted traits in the user like stereotypical behaviors. Both transcranial direct current stimulation and transcranial magnetic stimulation have also been used to enhance the capacity of working memory. It's hard to say how relevant this would be for most people, though. Transcranial magnetic stimulation is fairly expensive and might require ongoing use to have any benefit from it. Transcranial direct current stimulation (tDCS) is cheaper and could theoretically be done at home using a portable type device. However tDCS also has less selectivity in its ability to target brain regions, so it is unclear how useful it would be for the average person. Deep brain stimulation seems far too extreme for most people to gain any benefit from.

I think a major problem is how do you define "cognitive enhancement"? The brain is so interconnected that it is hard to affect one cognitive process without tangentially affecting another. You may be able to temporarily improve working memory, but how do you know this doesn't cause worsening on some other type of cognition? Even if you have a good knowledge of the brain, you may not be able to understand all the complex relationships between different brain regions and how they enable specific states of cognition to exist. It could be hard to manipulate things to a desired set point of functioning.

Another problem is that any single drug has the potential to induce massive changes throughout the brain. Even drugs that are rationally designed to be selective are not really all that selective. An SSRI is selective in the sense that it only affects a single transporter (serotonin). However it is non-selective in the sense that there are many serotonin receptors and an SSRI increases serotonin in the synapse thus causing increased activation of all these receptors. Also since you can't disentangle one neurotransmitter system from all the rest, a drug is basically going to change multiple different neurotransmitter systems at the same time. Serotonin regulates the firing of dopaminergic neurons in the ventral tegmental area for instance. SSRI's are of course, not considered cognitive enhancing drugs. However this rule applies to any pharmaceutical drugs that are in the pipeline for "cognitive enhancement".

Even if you use something like transcranial magnetic stimulation to target "selectively" the dorsolateral prefrontal cortex, this area has connections to many other brain regions. Activity in these other areas can be altered even if it is unintended. Also TMS either excites the neurons in the underlying tissue or inhibits their firing. However some neurons being activated may be excitatory, while some neurons being excited may be inhibitory. So your really getting a huge mix of effects that may difficult to really sort out. Some neurons being excited may cause other neurons to decrease their firing rate.

I kind of think that many people looking for cognitive enhancement may be somewhat disappointed currently. There are plenty of more far out technologies like neuromorphic brain implants or brain computer interfaces that could amp up cognitive processes dramatically. However these may be further away. I think most things for enhancing memory or attention now are very blunt tools with undesirable properties and may be of limited value for many.

Read more...

Virtual Fly Brain Computer Model

Most people think about ways that they can get rid of insects. However, some scientists are actually considering what it would take to create artificial insects with virtual brain's. Researchers are now planning to create a computer simulation of a fly's brain (drosophila). Could this virtual fly brain enable military mad scientists to fine tune a bug's functioning? Perhaps you could recalibrate a bug's pleasure circuitry so it would find enjoyment in injecting deadly poison into enemy combatants. Or maybe this might allow the development of increasingly complex forms of insect behavior like swarming or intelligence gathering by precisely altering the bug's neural wetware (with the help of the model).

I've previously mentioned about some attempts to construct computer simulations of the human brain (see computer brain simulation and blue brain). A human brain model in silico is quite a monumental task to undertake and may not come to fruition for quite some time. I've also noted before about my skepticism in the ability to model consciousness without the physics of our world. The human brain contains about 10^12 brain cells, 10^15 synapses and an exceedingly diverse array of synaptic proteins. At the very least, all of these are likely important for the overall functioning of the mind. Comparatively, the fly nervous system only has about 100,000 neurons. So it's surprising that modeling a bug's brain wasn't an obvious first choice. It's just so much simpler to do than a human brain. I personally find it highly probable that all insects have a simplified form of consciousness.

For the virtual drosophila brain, the researchers are proposing that sensory inputs and outputs could be added into the model. These senses include basically everything that would be part of a bugs perceptual experience (tactile, auditory, visual, gustatory, olfactory, even magnetosensory). An insect likely has a unitary consciousness that coalesces all sensations into one overall perception with discrete qualia. A more voluminous brain can probably enable a more complex conscious awareness. So we can assume that simpler organisms likely have a less complicated representation of objective reality. Specific "objects" may appear much cruder and far different to a fly than they would to a human being. An insect's consciousness only has to represent reality enough to drive behavior in a specific way.

A major stumbling block to modeling an insect brain is being able to scan all the relevant brain cell configurations and synaptic connections. They could visualize aspects of brain functioning using electron microscopy. However this would generate 26 terabytes of information and would require a huge amount of man hours to prepare the material. Recently faster methods have been developed to procure and analyze the data. This may not remain a constraint forever in the future.

The researchers mention that there are about 2000 to 3000 genes linked to human inherited diseases that are conserved between a fly and human. So being able to better understand the mind of a this insect has implications for human brain disorders as well. However, there are also considerable differences between the two, such as the fact that drosophila brain cell axons are not myelinated. This lack of myelination means that neural signaling may happen less quickly than in mammalian brains. Also drosophila doesn't really have the same blood flow and cardiovascular system as mammalian brains do. So oxygen reaches the neurons in a fly in a completely different way (not through red blood cells).

The researchers conceptualize the model as being used to predict the subsequent behavioral fly output when a specific neuronal adjustment is undertaken. Like how would a fly act if scientists were to upregulate a single receptor protein in a key brain region by genetic engineering? In the past, researchers have been constantly modifying the genetic source code of drosophila. However, this virtual model could potentially exponentially increase the understanding of how neural changes encode for behavior. It might become much easier to remodel a bug's functioning to essentially do whatever a scientist wanted it to do. They would basically be using a control type theory for the virtual model that could continuously be adjusted in order to refine it. So the virtual model may not necessarily have to be conscious in order to successfully predict the behavioral output of any brain change.

If successful, a virtual bug model could allow the creation of increasingly bizarre insect minds that have never before been seen in nature. Better modeling and understanding of these simulations might allow scientists to fashion more computationally efficient insect brains, for instance. Maybe they could increase the amount of proteins in the synapses or add myelination to the neurons so as to overclock the fly's brain. They might be able to tell ahead of time the subsequent effect on behavior. Studying the nature of consciousness would also be a fascinating aspect to this. Could you create a bug that was blissed out and euphoric? Darwinian natural selection usually precludes these extreme states of well being, but that doesn't mean you can't engineer them in an insect. Just imagine all the unique modes of consciousness that are possible by tinkering with a bug's genetic code. Perhaps these models could go a long way in helping to understand consciousness in an objective fashion and how it relates to behavior.

Obviously this project could take quite some time, but I think it could potentially happen faster than making a human brain computer model. The first models would probably be simplified, but could increase in detail over time. The fact that researchers can genetically engineer and breed flies so quickly also means that it would be much easier to test out the model to see if it was predicting behavior. A lot of interesting things could become possible in the future as this field matures.

Armstrong, J., & van Hemert, J. (2009). Towards a virtual fly brain Philosophical Transactions of the Royal Society A: Mathematical, Physical and Engineering Sciences, 367 (1896), 2387-2397 DOI: 10.1098/rsta.2008.0308

Read more...

Magnifying Taste Pleasure by Neuromodulation

I've been thinking about the nature of consciousness. This post will be a speculative and somewhat unrealistic "what if?" propositon about extreme qualia. Qualia is philosophical speak for "what it's likeness" (redness, tastiness, pain, etc.). Qualia is essentially how we consciously perceive aspects of the objective world through our senses. Will people be able to increase their capacity to enjoy foods? Will future neuromodulation techniques allow us to make the food we consume amazingly delicious? One current way of making a delectable meal is to slave hours over the stove using only the finest ingredients available. A future method to create a sumptuous mouth-watering taste experience may be done by using advanced brain manipulation techniques so as to actually enhance the mind's ability to assess pleasant foods. We can also possibly ensure that we are able to enjoy a diverse and unparalleled array of textures that exceed our current palate exponentially.

Using neuro alteration, one can potentially make any food taste exceedingly wonderful. These types of of extreme piquant qualia have previously been inaccessible to any conscious mind in the history of the earth. Darwinian natural selection precludes their existence. However using neuroengineering, the taste of the ambrosial food of the divine can be summoned up on command. Perhaps this marvelous savory taste perception can be applied to healthier foods so we don't become excessive unhealthy gourmands. Or maybe these sensations could be turned on without the actual consumption of food. A brain chip might enable us to elicit these feelings whenever a person so desired. The chip could be specifically designed to artificially induce extreme taste qualia that would be uncoupled from any sort of eating.

Scientists are increasingly unraveling the neural correlates of taste hedonics. Kent Berridge has done extensive research in determining the location of specific "hedonic hotspots" that are involved in sensory taste pleasure. The brain regions related to experiencing these phenomenon include the ventral pallidum and also the nucleus accumbens. Amping up mu-opioid and endocannabinoid receptor activation in these discrete brain areas has the capacity to make certain food much more appetizing. A lot of this research has been done on rats by observing the changes in their facial expressions when neurochemicals in specific regions are altered. They can actually tell that the rat is enjoying sweet food more by the way it licks its lips. I think for most people it should not be too surprising that drugs which affect the opioid or cannibinoid neurotransmitter systems have the ability to make food subjectively more palatable. People often use alcohol to increase food enjoyment and alcohol perturbs the opioid system. A pizza that normally might taste like fodder could taste amazing after taking a large amount of alcohol. Marijuana (a cannabinoid) tends to give people the munchies as does heroin (an opioid). These neurochemicals are dissociable from dopamine, which does not seem to increase "liking" of food. In the future, more sophisticated techniques may become available for precise taste perception alteration.

Taste hedonics has been an important evolutionary driver of behavior for an organism. On one level, the tongue has evolved the capacity to detect certain molecular configurations of matter. On another level the brain has evolved a way to assign value to specific types of matter that are placed in the mouth. Matter that has an overall beneficial affect on the functioning of an organism often subjectively tastes really good. Conversely, matter that is harmful to the organism may taste awful and is in general quite noxious. Sugar tastes sweet because this specific sensory qualia was adaptive, evolutionary wise, for our ancestors. Organisms that found sugar to taste sweet did better than organisms that found sugar to be repugnant or neutral in flavor.

Currently, there has been a substantial environmental paradigm shift for humans. The advent of the supermarket has made a plethora of sugary foodstuff items available in plenitude. This has happened while the neural signatures for assessing sweets in the brain has remained largely the same. This may not be the case forever, though. In the present/future, evolution will likely select against finding sugary foods as being too sweet. A person who consumes a lot of sugar probably ends up overweight and thus reduces their reproductive potential overall. They may be less likely to have children. Rational brain engineering may allow us to overcome this negative effect, however.

We can assume that different animals may experience somewhat similar forms of taste hedonics. All mammals may enjoy foods with high glucose concentrations for instance. However a specific item that tastes scrumptious to one organism does not necessarily taste good in the same way to another organism. On the earth there are trillions of conscious minds and each likely experiences some sort of perceptual taste hedonics. Consuming food is integral to the life of almost any animal. Each conscious brain is the result of a unique aggregation of atoms that encode for a particular perceptual experience. Will we be able to categorize and actually understand these perceptual qualia of all organisms on the tree of life and their neural correlates? Once we have correlated these neural signatures, then can we possibly replicate it in our own brain chemistry and amplify those textures of experience to dizzying heights?

Why don't these extreme qualia currently exist? Suppose you were to saturate a single macroscopic item (example sugar) with extreme positive value. Imagine if a neural pathway evolved to find this specific foodstuff item exceedingly delicious. This item would taste so good that you could never get enough of it. This foodstuff would in essence become an all consuming driver of behavior. A single item would become a veritable black hole abyss by which an individual's behavior would be inescapably and indefinitely drawn toward. In the evolutionary fitness landscape this would be maladaptive. Evolution normally wants an organism to have a diversity of behavior (like eating a variety of food items that serve separate biological purposes). Extreme qualia drives an organisms behavior toward that one item and would potentially cause them to ignore other perhaps equally important items in the environment. Also sometimes evolution just does the least amount of work necessary to motivate an animal. Why make something extremely tasty when you can get the same behavioral output by making it just taste pretty good. Evolution often tends to spurn excess.

Can you engineer extreme qualia while at the same time maintaining a diversity of behavior? Also, can we disentangle these qualia from the actual act of consuming food? How do you ensure that a person's behavior does not get stuck in a suboptimal rut? Perhaps these future extreme qualia will serve no functional purpose. Maybe they will merely be based on whatever arbitrary whim a person has. Like activating and deactivating a brain chip whenever they want to experience a specific sensation. I think the future will become extremely interesting if we can potentially affect discrete qualia and shape them to whatever we desire. Neuroengineering may eventually allow extreme qualia that are currently closed off to current brain wetware.

Read more...

LILFU and the Brain

Technology Review has a new article about brain stimulation using ultrasound. Basically it talks about some of the stuff that I have previously mentioned, but there are a few new things about more recent experiments done. Here is an excerpt;

"With ultrasound, we have a much better spatial focus than [with] DBS," says Tyler. "And unlike TMS, we can get anywhere in the brain." Ultrasound--consisting of sound waves with a frequency above 20 kilohertz--has been used for decades in medicine to image muscle, organs, and fetuses.

Recently, ultrasound has been shown to alter the morphology of neurons. So it should be interesting to see what sort of manipulations may become possible in the future. Ultrasound could potentially be used to fine tune several aspects of brain cell functioning.

Deep brain stimulation has been shown to spark new neuron growth in key brain regions. Research has shown that these neurons are functional. So I think researchers may find ways of using non-invasive brain stimulation (deep TMS or possibly ultrasound) to do the same thing.

Read more...

Neuro Brain Thermodynamics

The brain is a metabolically expensive organ that uses quite a bit of energy. It's no surprise that it also generates a decent amount of heat during this energy usage process. A new paper has come out that poses and answers the question as to whether there is a thermodynamic limit to brain size (evolutionary wise). The author is basically asking how big can a brain get before it becomes too hot to function properly? What sort of constraints does evolution have in constructing a bigger brain, given the laws of our universe? The answer he gives, in short, is that there is plenty of room thermodynamically to evolve a larger brain.

The paper first discusses the main cause of the generation of heat in the brain. The Na+/K+ pump helps to maintain the cell membrane potential of every neuron. The pump allows a neuron to have a high concentration of K+ ions and a low concentration of Na+ ions inside of it. The protein molecule pump hydrolyzes an ATP molecule, using that energy to move 3 Na+ ions out of the neuron and 2 K+ ions into the neuron. These ions are integral to the electrochemical signaling of a brain cell and their concentrations change in response to the propagation of an action potential down a neural axon. So the pump is necessary to restore ion balance to a neuron after it fires.

The whole paper is a bit mathematically intense. The author's intent is to figure out how much energy an aggregate number of neurons use, mostly focusing on the Na+K+ pump and neglecting brain glucose utilization. The author goes on to discuss how the brain regulates the heat generated from all of this work that occurs and the specific constraints on neural processes. Some heat radiates from the scalp. Cerebral blood flow (cbf) is another method that the brain uses to cool itself. Up the evolutionary mammalian line, cbf essentially scales with brain volume. So I think that means the amount of blood vessels are basically proportional in creatures that have varying brain sizes. Due to this scaling up, there is apparently a small decrease in the rate of blood flow as you go from simpler to more complex organisms (mouse to human for example), because of more surface area coverage.

Increasing cbf in a specific deep brain region can cause a resultant decrease in brain temperature there. According to the author mammal's brains can sustain a temperature upwards of 42 Celsius without becoming damaged. Though, the optimum temperature may be much lower than that. Certain drugs have the ability to increase the brain's temperature in part by vasoconstriction. Cerebral blood flow only acts as a coolant inside deeper brain regions where the blood is cooler than the surrounding brain tissue. So it's really not a coolant in the same manner as that in a heat engine. The cbf actually heats up more superficial brain regions that are closer to the skull so the mechanism is not uniform. The main purpose of cerebral blood flow is to bring glucose to neurons for their basic energy need. So the cooling ability is sort of a secondary aspect of blood flow and is probably not ideally suited for that purpose. Evolution has basically co-opted one process for a different purpose entirely. The blood flow's ability to cool is more important for larger brained mammals and less relevant for smaller brained ones where heat can dissipate from the head easier.

The author of the paper notes some of the constraints of the brain taking into consideration excess heat production. He suggests that thinner axons/dendrites result in excess heat. However, he estimates that the axon's diameters are at a magnitude higher (averaging 12-1500 times) than the lower bound diameter that would be problematic. He also talks about the heat bounds relating to the propagation of neural signals and density of axonal packing. The author concludes that deep brain temperature is only weakly correlated with brain volume. So the brain could easily be scaled above the 5 kg limit of current land mammals. However, that is assuming that no other methods are utilized by evolution to "overclock" specific brain regions.

I think that evolution finds whatever way it can to increase the brain's computational capacity. I've mentioned previously about some of the possible ways. Evolution exploits any pathway easily available. The firing speed of neurons is an aspect of overall computational capacity. However there are limits to this facet of brain functioning. Neurons can only fire continuously so many times before the sodium concentration in the cell becomes too high. This is especially true if neuronal axons were to become too thin. Those Na+/K+ pumps can only pump sodium ions out of the neuron so fast in certain cases. Evolution can possibly add more pumps, but then that requires more energy which potentially generates more heat. The pumps work relatively slow, so even a maximum amount of them might not overcome a certain limit. Eventually you might run up into an insurmountable wall with this attribute of brain function. So evolution may have to do something else, like increase the overall amount of neurons.

Sometimes the path that evolution follows is unexpected. Average neuronal firing rates in larger brains are actually less frequent than that of neurons in smaller brains. So there may be some limitation to increasing neuronal firing rate over the course of evolution as you scale upwards in size. You would also think that evolution would first do something like maximizing the amount of proteins in the synapse before it went on to increase overall neuron count. However evolution tends to fill in some of these finer details later on, instead of in a logical linear fashion. Scientists will probably increasingly figure out why occurs as time goes on. The author of the paper does his part to elucidate a few of these possible constraints on how the brain is arranged.

With increasing brain complexity there is the potential for more pathways to open up that can be exploited to further increase overall computational capacity. With more complexity, however, there is also the possibility for evolution to have a harder time navigating a proper way forward. There may be too many entangled systems whereby changing the variable of one thing could have a negative effect on something else. When cerebral blood flow is too high, for instance, it can damage the brain. So evolution may not be able to just take obvious routes (like increasing brain blood flow speed) to decrease brain heat. Not to mention how blood flow is mainly involved with delivering energy to cells. So any change in cbf could potentially negatively affect that process as well. Short of developing a whole secondary cooling system, evolution is stuck co-opting cbf as a coolant system. Also more blood flow in the brain may mean less room for computational purposes. Of course, this paper indicates that increased cooling might not be necessary. It may tend to get difficult as evolution moves forward to undo things or reach a path that is radically different from what previously evolved. Evolution has to essentially make due with whatever it has.

In the past, scientists from Japan have constructed a "heat pipe" that can be implanted directly into discrete brain regions. This device can essentially be used to cool brain areas by diverting heat to an outside heat sink. The researchers developed this implant specifically for the purposes of reducing epileptic seizures. Over-excited neurons in an epileptic seizure are more active, use more energy and thus generate more heat. This excess heat that is generated can cause a feedback loop exciting more neurons to fire thus potentially prolonging the seizure. So cooling the brain could be a way of reducing problems associated with this condition. It is possible that this could be used to better regulate the temperature of future engineered brains.

I think the fact that our universe allows brains to evolve to the size they do may be an example of the anthropic principle. First, there is the rare earth perspective of life. Our planet is situated in a near perfect distance from the sun and is neither too hot nor too cold to sustain life. The gravitational pull of our planet may be at nearly the right level that allows a larger brain to evolve. Also our solar system is located far enough from the center of the galaxy so as to avoid excess radiation. These are only a few examples of our special situation in our own galaxy/universe. String theory predicts that our universe is merely one region out of a larger multiverse. In the multiverse there are different vacua that may have varying constants. A majority universes in the multiverse may not be able to sustain any life at all. Some universes may contain selfish replicators, but they never be able to evolve a nervous system or the capacity for sentience. Perhaps in an even smaller subset of vacua in the multiverse, a brain/nervous system can evolve, but maybe it cannot attain a complexity greater than that of a mouse's brain or an insect's or even less. There might be too many design constraints inherent to that specific universe for it to get any bigger. The physics of our own universe is perfectly suited to developing a human level intelligence. The heat capacity of water may allow the brain to maintain a fairly stable temperature, for instance. Definitely quite amazing when you think about. Perhaps we will eventually figure out everything related to this more speculative science in the future.

Karbowski, J. (2009). Thermodynamic constraints on neural dimensions, firing rates, brain temperature and size Journal of Computational Neuroscience DOI: 10.1007/s10827-009-0153-7

Read more...

Brain Function and Ultrasound

Researchers are using GPU processors to enhance the imaging capacity of ultrasound. (see GPU engine enhances ultrasound-detected brain motion calculations). They are specifically using an ultrasound transducer array (shown left) positioned around the head to detect brain tissue micropulsations. Obviously more computing power allows for better analyzing capability of ultrasound pulses. In the past other researchers have been working on better focusing of ultrasound, which could further enhance the images created. I've mentioned previously about some of the other intriguing applications of ultrasound. This new research shows a few interesting trends that are developing. I think a lot of the hard work has already been accomplished in successfully getting ultrasound pulses to focus on brain regions. So ultrasonic brain stimulation may become a simpler undertaking. It is basically riding on the back of all this other research that has already been done. Throw in a dash of an accelerating technological growth rate and the ability to manipulate the brain could increase quite dramatically and cheaply.

Could you have something worn on the head (like the device above) that would stimulate brain regions selectively? In theory, I don't see why not. 20 years down the line you may potentially be able to buy a neural stimulation device based on this principle. This may be technically possible, but I'm not sure how it could be done in a safe manner. It has been mentioned on other blogs that there is a site that seeks to build a home transcranial magnetic stimulation device (OpenStim). According to that page, the OpenStim project has been put on hold due to inactivity. At least one person in a forum has discussed about trying to use ultrasound for brain stimulation purposes. It seems premature to be doing something like this now when so little is known about its effects on brain chemistry. Maybe in the future a community will crop up attempting to build an "open source" ultrasonic neuromodulation device.

I'm relatively libertarian when it comes to a lot of mind alteration (at least when it comes to drugs). People have been manipulating their brain chemistry via the use of psychochemicals for quite some time. However this sort of brain stimulation seems like a different animal. It really requires a knowledge of what areas to target in the brain for a desired effect, which I think may be above most people. Maybe intuitive computer interfaces will allow people to offload the hard work to their PC. I do like the idea of giving people some autonomy over how they want to manipulate their brain. Open source brain manipulation coupled with the sharing of neural pathways associated with specific emotional states (using brain imaging and the internet) definitely sounds like an appealing future society. Perhaps with robust safety mechanisms in place, this could become a reality. With any brain alteration you have the potential to affect the balance of power between two individuals. Is it done for the good of the individual, or done for the good of society (i.e. other people)? What would a person use this brain stimulation device to accomplish? Definitely some interesting questions come up that don't have easy answers.

Read more...

Brain Synapse Computational Capacity

Researchers are uncovering another layer of complexity as to how the brain functions. Brain cells communicate with one another by chemicals through synaptic connections. The human brain contains billions of neurons and each neuron has a large amount of synaptic connections to other neurons. Each synapse itself contains a variety of receptor proteins that can alter the gross firing pattern of a neuron. It has only been recently that scientists have been able to better understand the role of synaptic protein interactions in the computational capacity of the brain. A lot of this activity functions at an even lower level than overall neuronal firing.


In the past, researchers have found that different organisms on the tree of life have varying amounts of these receptor proteins in the individual synapses of their neurons. As you go from simpler organisms up to mice, there are an increasing number of synaptic molecules. In that past study, the scientists had investigated approximately 651 different genes that directly encode for proteins in the postsynaptic junctions of mouse neurons. They specifically focused on proteins that can be phosphorylated. Phosphorylation of protein molecules changes their functioning. The researchers looked for the same proteins in a variety of other life forms besides mice that had varying levels of complexity (invertebrates, non-mammalian vertebrates and other mammals). Lower complexity organisms like invertebrates had about 45% fewer of these synaptic proteins than the mouse synapse, while an even simpler organism like yeast only had about 21% of the number of proteins.

Now those very same researchers have published more work uncovering the complex interactions of the molecular proteins in an individual synaptic connection.

Here, we define maps of molecular circuitry within the PSD based on phosphorylation of postsynaptic proteins. Activation of a single neurotransmitter receptor, the N-methyl-D-aspartate receptor (NMDAR), changed the phosphorylation status of 127 proteins. Stimulation of ionotropic and metabotropic glutamate receptors and dopamine receptors activated overlapping networks with distinct combinatorial phosphorylation signatures. Using peptide array technology, we identified specific phosphorylation motifs and switching mechanisms responsible for the integration of neurotransmitter receptor pathways and their coordination of multiple substrates in these networks. These combinatorial networks confer high information-processing capacity and functional diversity on synapses, and their elucidation may provide new insights into disease mechanisms and new opportunities for drug discovery.

The researchers of this specific work used proteomic and also computational methods to disentangle all the relationships between these synaptic proteins. So being able to determine these relationships is really in some respects on outgrowth of certain accelerating trends in computing power and protein/genetic analyzing capability. I mentioned previously that Henry Markram talked about how certain newly developed methods are vastly speeding up scientific research. Research in uncovering some of these molecular mechanisms has moved rather slowly in the past, but greater computing power and software analyzing capability has the capacity to greatly accelerate progress.

The researchers have found that all of these molecular networks in the synapse may underly some of the overall computational capacity of the brain. You can read more about it at the press release here.

The team's discoveries led researchers to the conclusion that the brain is organised like the internet, where billions of these molecular computers - intricately complex in themselves - are connected by billions of nerve cells.

So evolution has exploited multiple avenues to increase the brain's computational capacity. The avenues that were taken exist at differing "levels". Overall brain cell number is a "higher level" avenue. The human brain contains many more neurons than that of a mouse and other lower level organisms. Evolution has also favored mutations that cause increased branching and growth of neuronal axons. Mutations which increase levels of glucosylceramide in the brain, for instance, can increase the amount of neural axon terminals. There is evidence that recent evolutionary selection pressure on humans has favored mutations which alter the amount of glucosylceramide and that these specific mutations may lead to a higher intelligence. More axon terminals equal more synapses connecting each neuron. At a molecular "lower level", evolution has favored increasing the number of proteins in each individual synapse and a more complex interaction between those proteins. There are other potential ways that evolution may have worked on as well, which I won't mention here.

By merely simulating a higher level of brain functioning (overall neuron firing/activity) on a computer, researchers may totally miss a substantial amount of lower level functioning. So future computer brain simulations will likely have to model all of these protein interactions to function in a manner similar to a real brain. Even then, it is not clear if they will be successful in modeling the mind exactly (especially without the underlying physics of our world). I think you can probably model aspects of brain functioning very well on a computer, even with a simplified model (like without molecular interactions). However, getting a computer simulation of an entire brain to function exactly like a real brain (meaning it would have consciousness), may be a difficult task if not an impossible one. This new research, though, will certainly have an impact on our understanding of how the brain functions.

Read more...

Encephalon 69th Edition

Welcome to the 69th edition of the Encephalon blog carnival. This carnival is devoted to presenting blog posts that cover neuroscience and psychology related topics.

First up, Neuroanthropology does an in depth analysis about the twitter phenomenon in "Fear of Twitter: technophobia part 2". I personally have not been inclined to get on the twitter bandwagon. Do I have twitter phobia? Perhaps. Not sure if I'm interested in another time waster. A couple of topics in that article include a discussion about matters related to privacy and twitter's potential affect on cognitive processes.

Brain Blogger has a post about "Free Will and the Philosophy of Science". The author talks about the relationship between neuroscience and scientific determinism. There is further discussion about the topic in the comment section too. There is a little bit of confusion on the part of the blog post author and some of the commenters about whether the physics of our world is deterministic or not. The most widely accepted interpretation of quantum mechanics among physicists (many worlds) is deterministic. So does a deterministic world deny the existence of free will? Well I think that depends on how you define the term "free will". As mentioned by the blog post author in one of the comments, the term "free will" can be rather vague. So I think the discussion won't be particularly fruitful if you don't properly define what exactly "free will" is. It's a topic I myself might explore in a future post.

Sharp Brains writes on the subject of whether taking art classes can boost academic achievement in other fields in "Arts and Smarts: Test Scores and Cognitive Development". There is conflicting data as to whether participating in the arts is beneficial for test scores in more conventional domains (like math or reading ability). There is also the issue of whether correlation proves causation in some of the studies.

Neuroanthropology has another lengthy article that discusses about a "neuro" term in "Who you callin’ a ‘neuroconstructivist’?!". Another "neuro" word? What does neuroconstructivist mean? You'll have to read the post if you want to find out the whole story. It looks like part of being a neuroconstructivist entails taking environmental effects on the brain and its development more seriously. The expression of brain genes can be dependent on social environmental factors, for instance (i.e. social interactions). From my own reading of the Neuroanthropology blog, I tend to think they sometimes overstate the case for certain environmental social factors being able to shape an individual's brain to a desired output. Nonetheless, they usually have well reasoned posts. I also think it is important to stress that while the environment can have an impact on brain functioning, brain functioning itself can impact how we consciously perceive the environment. Changing brain chemistry (like by using a neuro drug as one extreme example) can alter our perception of reality (i.e. the environment) and this can affect all subsequent brain/environment interactions as a result of that altered perception.

Brain Blogger has another post discussing new objective testing methods for Alzhiemer's disease. Volumetric brain imaging analysis of the hippocampus can be used as a biomarker for Alzhiemer's disease. It's unfortunate that there aren't too many therapies that are truly effective in treating/preventing Alzhiemer's even with earlier detection. A lot of the new neurotechnologies discussed on my own blog will probably have only a limited effect on the disease process too. I also think it will be getting harder and harder to actually get FDA approval for many new neurological therapies in the future. The current regulatory environment is not conducive to this sort of thing and it will probably only get worse. Hopefully my pessimism will prove wrong.

Cognitive Daily has a post entitled "How wrong is it to use a kitten for personal ughhh... pleasure?". Very embarrassing post title to describe. It deals with one particular sort of kitty induced pleasure. Do we really need to be giving people these sorts of ideas? Aren't there enough weirdos out there who will try anything? The post talks about how people have a tendency of rating things as being more immoral when they are exposed to something else disgusting (fart smell, disgusting movie) that is unrelated to the poll. Conversely you can make people judge something as being more "moral" as well by having them do certain things beforehand (having them wash their hands, for instance).

Sharp Brains also mentions about a Cognitive Health Track at Games for Health Conference.

Well, that's it for now. The next Encephalon installment will be hosted at TBD, May 11th, 2009. More information about future editions of this blog carnival and details about making post submissions can be found at the Encephalon archive.

Read more...

Brain Technology

A talk about the Blue Brain project has recently been given at the European Future Technologies Conferfence. You can read some more stuff about it at this german site (they have an mp3 audio at that site, but it's in German unfortunately). Blue Brain seeks to simulate the functioning of the brain via a computer. Henry Markram is the main researcher on this project is. Markram is the founder of the Brain Mind Institute (BMI) and he has made some key insights about the functioning of neurons in the past.

Currently the researchers have completed the first phase of the project. It sounds like they still have only simulated 10,000 neurons and 30,000 synapses of a rat cortical column. The researchers are integrating the simulated brain into a virtual reality agent. So they will have a simulated animal that will be able to function in a virtual reality environment. This will allow researchers to view the changes in neuron functioning as the animal moves around its virtual environment. Markram told the conference "It starts to learn things and starts to remember things. We can actually see when it retrieves a memory, and where they retrieved it from because we can trace back every activity of every molecule, every cell, every connection and see how the memory was formed." It's probably too soon to say how successful this will be. I think it may be a stretch to say that they can actually create a brain with the properties of consciousness. So I'd be a little skeptical, but Markram probably knows more about this than I do.

In future phases of the project they will use faster supercomputers that will allow scientists to add more details to the simulation. They plan on simulating biomolecular pathways and also gene expression patterns. Here's an excerpt from Markram's keynote speech at the beginning of the conference (Shaping 21st Century Science and Society) that sounds very similar to Ray Kurzweil in many respects.

There is no sign that the exponential growth in computing power is decreasing which means that we will see a 1000 fold increase each decade (exascale (10^18) by 2020, zettascale (10^21) by 2030, yottascale (10^24) by 2040). Such a growth in computing power within our lifetime comes with extreme challenges where information and computational devices need to be energy-efficient, fault-tolerant, capable of self-repair, and where storing and processing of the vast volumes of information generated is carried out with novel automated and highly intelligent information processing systems. The exponential trends of minimizing size and cost and maximizing speed and efficiency are so extreme that the evolutions and revolutions in ICT over the next 10-20 years will be equivalent to those of the past 100 years.

Markram talks about 3 elements that are transforming science. He mentions that data acquisition is increasingly becoming automated. He says there has been an exponential growth in the amount of DNA being sequenced and he predicts that we will know the DNA sequence of all organisms on the planet in 3 decades. There is also a greater ability to process large amounts of data through the use of statistical correlations and machine learning. He also says the computerization of science is leading to increased collaboration and productivity.

I think it seems feasible that these computer simulations will eventually allow extremely personalized medicine. Brain scanning technology is continuing to get better, perhaps down the the molecular level. In the future a person may be able to get their brain scanned and then have a computer simulation of their brain on their home computer. Ray Kurzweil has talked a lot about these exponential technological trends and how they could transform what it means to be human. Will there be ways to increase human happiness or human intelligence exponentially in the future? Who knows. I think a lot of interesting things will become possible assuming these simulations become increasingly accurate and realistic in their portrayal of the human brain.

Read more...

Artificial Brain

I found a blog that has some more information about the DARPA SYNAPSE project. The SYNAPSE project seeks to create a neuromorphic artificial brain. That blog also has a bunch of older posts that specifically relate to this project. Will this actually lead to anything? I suppose it's always possible, but I guess I'm fairly skeptical that it will accomplish much.

SyNAPSE is a complex, multi-faceted project, but traces its roots to two fundamental problems. First, traditional algorithms perform poorly in the complex, real-world environments that biological agents thrive. Biological computation, in contrast, is highly distributed and deeply data-intensive. Second, traditional microprocessors are extremely inefficient at executing highly distributed, data-intensive algorithms. SyNAPSE seeks both to advance the state-of-the-art in biological algorithms and to develop a new generation of nanotechnology necessary for the efficient implementation of those algorithms.

Another blog also has some information that pertains to neuromorphic electronics.

“Neuromorphic engineering takes inspiration from the signal processing structures found in the brain and physical attributes of animals to design new computers and robots capable of the amazing sensorimotor feats seen in nature. From neurons to behavior, the low-power, robust, real-time, and adaptive nature of biological systems serves as a proof-of-concept of the unique implementation developed by evolution. These principles have been applied to software models of sensory processing, VLSI implementations of neural circuits, and robot design.”

Also I found this interesting paper called "Framework and implications of virtual neurorobotics" (PDF) that discusses about using virtual reality environments to further the development of neuromorphic electronics.

More recently, investigators are focusing on the core assumptions of the brain “algorithm” itself—trying to replicate uniquely “neuromorphic” dynamics such as action potential spiking and synaptic learning. Only now are large-scale neuromorphic models becoming feasible, due to the availability of powerful supercomputers and an expanding supply of parameters derived from research into the brain’s interdependent electrophysiological, metabolomic and genomic networks. Personal computer technology has also led to the acceptance of computer-generated humanoid images, or “avatars”, to represent intelligent actors in virtual realities. In a recent paper, we proposed a method of virtual neurorobotics (VNR) in which the approaches above (social-emotional robotics, neuromorphic brain architectures, and virtual reality projection) are hybridized to rapidly forward-engineer and develop increasingly complex, intrinsically intelligent systems. In this paper, we synthesize our research and related work in the fi eld and provide a framework for VNR, with wider implications for research and practical applications.

I don't see much discussion about the ethics of doing this type of research. I'm not sure if they are attempting to make artificial brains that are conscious or not. Bringing about a conscious AI could have a variety of ethical implications. Most likely, though, progress in this field will be fairly slow. I wouldn't expect much to happen for quite some time, if at all.

Read more...