Brain Science Podcast discussion

Controversial science of brain imaging
(Scientific American Blog, July 5 2012)

"It seems that neuroscience reductionism is now replacing its genetic counterpart to find an explanation for everything about ourselves using neuronal correlates instead of genes. While I understand and acknowledge the astonishing advancement of science and its benefits, I tend to think that this infinite spiral of reductionist path may not allow us to make the leap of truly understanding who we are.

Science has already come a long way helping us in understanding the nature. Perhaps being creatures of nature, some aspects of it will forever elude us and we may never fully grasp the human experience."


http://blogs.scientificamerican.com/g...

Having just re-listened to episode 83, I feel very educated by Dr. William Uttal's explanation for why fMRI imaging does NOT deserve all the press it gets.

I agree that it's over reductionist. Having Uttal's critiques fresh in mind had me reading this Scientific American blog with much more understanding.

Neuroimaging is our generation's phrenology!

There is something to be said for neuroimaging in diagnostics, but as a research tool, it has severe limitations (even the newer diffsuion tensor tomography-makes pretty pictures though). And when it comes to reductionism, it just doesn't work for understanding the operation of complex systems (reductionism tells us a lot about very little). However, reductionism in neuroscience does provide some clues to guide our pursuits if we shake off some of the historical dogma that perpetuates misdirection due to popular assumptions. I think it's really a matter of asking the right questions and re-evaluating a lot of interpretations in light of exceptions.

Dalton, should neuroimaging not be used in research? Isn't a little information better than none?

Re: reductionism--can you give examples of "historical dogma that perpetuates misdirection due to popular assumptions"?

Mitchell wrote:
Re: reductionism--can you give examples of "historical dogma that perpetuates misdirection due to..."

Mitchell, if I may insert myself into the conversation, one possible example is the idea that dietary fat, especially saturated fat, causes disease and obesity, and that obesity is a problem of the will (psychological) caused by overeating and/or underactivity.

According to the book Good Calories, Bad Calories by Gary Taubes, all that is a myth and the evidence found by studies tends to refute it. However, because of historical dogma, even questioning the assumption (i.e. not starting with the "dietary fat => body fat" ideas as a first principle) is verboten.

[I haven't read the book myself, and I don't (necessarily) subscribe to that view (yet - we'll see once I read it). However, it's on my list, and I've read (in part, in reviews here on Goodreads) that Taubes makes a big effort to support his conclusions with tons of publicly-available research.]

Another example, more close to home, is neuroplasticity - it was flatly denied by most neuroscientists until very recently, when they could no longer deny the preponderance of evidence.

How about examples re: reductionism? And in any case, if there are examples of dogma perpetuated by popular assumptions about reductionism, does that invalidate reductionism as an scientific methodology?

Mitchell, neuroimaging has its uses in diagnostics but the latency of response and lack of detail don't make it all that useful for research.

Examples of reductionist dogma: Ok, I'll go out on my limb here and tell you what I think.

The first one has to do with the notion that the more synapses are used, the more strength they gain - an assumption because of increased size. However, there is another way to interpret the relationship to synaptic size and the frequency of it's use and that's to increase the surface area to accommodate more vesicles in order to deal with increases in duty cycle.

The next bit of dogma has to do with the spiking behavior a neuron exhibits. The notion that some form of coding exists within the spike train beyond intensity of response makes no sense either for no one has yet accounted for the corresponding intracellular decoder in the target neuron. Another interpretation for such spiking behavior is simply that the neuron is behaving like an electrochemical cell employing capacitative or metal oxide varistor like structures who's active synaptic fields are collapsing serially producing a spike train.

The next example (in my opinion) is the popular notion and statement that the synapse is the location for a memory. Nope, synapses are selectors for cells in a hierarchical structure that leads to a cell representing a memory which, together with feedback (or as Edelman would refer to it – re-entry), plays back the feature collections representing a memory through earlier stages in the hierarchy. I know that the Grandmother Cell hypothesis is rapidly loosing ground but it hasn’t been replaced with anything as yet. I suspect the biggest problem the Grandmother Cell hypothesis has is the name. If you view the pyramidal cell as a hub cell (what else can it be with 5K to 20K connections each), then it’s very plausible.

In regards to distributed networking, the alternate interpretations seem to make a better fit for complex operation in a competitive environment.

These are both reductionistic examples but they don't get you past the single cell stage. Understanding how the brain operates requires the consideration of how thousands and millions of cells interact competitively in a multidimensional collection of distributed networks. That's why I said, reductionism tells us a lot about very little.

Just for fun, here's one last one to ponder. In reality, it's not the neurotransmitter molecule that imparts some magical effect on brain operation, it merely serves to make the signal connection between cells. So, one has to wonder why nature has produced such a variety of neurotransmitters. Having pondered this and not seen it else where, but very interested in connectivity, I took the view that they may play more than one role in the overall development of the brain. Some how, as the neurula stage progresses and generates cells, they eventually have to connect up in a pattern that will kick start the brain successfully. The diversity of synaptic types and related neurotransmitter together with the cillipodia tipped axons growing and sampling the environment implies a chemical attractant, and what would be more likely than the diffusion gradient of neurotransmitters to ensure that the necessary connections get made under the influence of natural physical and logical constraints.

Sorry, Dalton, but I don't understand how this is reductionist dogma as opposed to neurophysiological dogma. If I understand your point correctly, you're not disputing Hebbian dynamics, only the mechanism that drives it. Similarly, memory may not be maintained in the synapse. I recall a BSP episode where glial cells are thought to have more of a role than just "nurse" cells and may be involved in memory. I like your suggestion that neurotransmitters have multiple functions beyond simple synaptic jumpers.

But according to Wikipedia, reductionism is an approach to understanding the nature of complex things by reducing them to the interactions of their parts, or to simpler or more fundamental things. Yes, understanding the brain requires "consideration of how thousands and millions of cells interact competitively in a multidimensional collection of distributed networks," but that doesn't negate the need to understand the system at the cellular and chemical level. Indeed, your theorizing addresses these levels. Sure, we're taking baby steps with Aplysia and C. Elegans when we map their neural networks, especially compared to the massive complexity of the human brain, but we can't run before we can walk. It's not that reductionist dogma is the problem; just that the current dogma is inconclusive. That's science all over.

I’m sorry Mitchell, I guess I didn’t explain the basis of my statement, “reductionist dogma,” very well and why I think a reductionist approach can’t provide an understanding of how the brain works. I don’t mean to negate the contribution of reductionism for establishing some of the fundamentals that guide understanding of the larger problem. What I mean is that (in terms of how the brain operates), it’s a “Can’t see the forest for the trees” problem. It’s nice to see that the trees have roots, but you need to step back to put the forest in perspective where you can discern the integrative and interactions between all the foliage. That’s the way complex system science is, reductionism only gets you so far and without the broad view, interpretations at the reductionist level is prone to misinterpretation. It’s commonly said that complex systems are unfathomable, yet are governed by very simple rules. I believe that’s because the claim is coming from the mathematicians who can’t as yet wedge that many variables into a formula – the dynamics unfold over time and look chaotic, but in reality, are deterministic. The key to understanding complex phenomena (I believe) is to first imagine what plausible functions could produce the phenomenon based on what reductionist approaches have revealed (with a skeptical eye to reveal misleading extrapolations). Both approaches have to fit together and not contradict each other. When they do contradict each other, it becomes possible to discover what the truth may be.

In my previous posting, I gave three examples of reductionist dogma (synaptic strength, spiking signals, and synaptic memory). There is no doubt that the physical characteristics of each are real and exist. The dogma is the popular interpretation of what each of those things mean in regards to brain operation and production of Mind. My criticism is that those engaged in reductionist neuroscience make claims without regard for how they fit at the system level. For anyone attempting to emulate the system, assuming that memory is located in the synapse, will take them in the wrong direction and doom the project to failure. In AI, it has been popular for apply weights to every connection between nodes, and as the result – COMBINATORIAL EXPLOSION. Take Watson, the supercomputer developed by IBM to challenge and win against Jeopardy champs, 32,000 CPUs, and it couldn’t even take questions in an audible format and used Bayesian Statistical approaches to data mining from 1000’s of terabytes of data. Yikes! It has been said that the electrical power consumption was that of a small city. Impractical for anything more than some way to get bragging rights and after OS/2, they need a boost. I’ve been following AI development since the mid 80’s and it’s really gone nowhere in terms of emulating a brain and general intelligence. At this time in my life, I really believe that conforming to the reductionist dogma is what has held them back. You constantly hear that we just need more powerful computers and that the mind will EMERGE. In my mind, emergence is another dogma. Yes, there are dynamically allocatable neural resources in the brain, but its development and connectivity are far more deterministic than the popular views claim. Deacon has noticed that constraints, naturally occurring and logical, impact what can be in a very big way and that we fail to take them into account. He refers to it as “what’s absent” and the failure to consider constraints is really quite natural. Constraints only get our attention when they block the way to success, however, brain development and connectivity rely on constraints to supplement genetic determinants.

So, as you can see, I’m quite opinionated about some of this stuff. Admittedly, I tend towards what in my mind is plausible and workable with respect to my objective. Sometimes it conflicts with popular theory and hypotheses, and sometimes, it just fills in where no theory exists. I’m always open to consideration of alternate and original suppositions.

First of all, I agree completely with you that "emergence is another dogma" and that the development and connectivity of neural resources are far more deterministic than the popular views claim. But unfortunately, the use of the term reductionism is as much a political issue as biological. There is a school of thought that uses a fallacious argument they call reductionism to try to generate philosophical opposition to the determinism of our genetics. These folks, some of whom have been interviewed on BSP, want to retain the dogma of free will in terms that resist or deny limitations inherent in our neural makeup. As an argument to invalidate the reductionist model they invoke fanciful strong emergence in which the results are greater than the sum of the parts. This accounts, they believe, in human exceptionalism, that humans are special creations that don't play by the same rules of biology as other animals. They want to believe, for instance, that human consciousness enables choice and control, but anyone who has been paying attention to BSP knows this is (mostly) wishful thinking.

I agree that it is inappropriate to make claims that are not supported by the science. Some researchers are certainly guilty of this, although in some cases, they are proposing theoretical, hypothetical implications, which drive future inquiry. In any case, that does not invalidate the scientific results. You mentioned diffusion tensor tomography, which surgeons use successfully for pre-operative planning for strokes and tumors. Not too shabby. You also get those pretty pictures as you pointed out, which show neural tracts in nice detail. It may not definitively tell us what connects to what, but it's a helluva start and should go into the evidence-based literature.

You said the physical characteristics that have been described for neuronal and synaptic interactions are real and exist, but in some cases, the meaning and significance of these interactions have been misinterpreted. I, too, have taken umbrage with theoreticians regarding what I consider to be unfounded positions. Dogma that is not supported by the evidence needs to be drop-kicked. However, it behooves us to differentiate between science-based and philosophy-based issues of reductionism. Personally, I don't feel that it's necessary to haul down the reductionist flag, but if you must, then we need to be clear that it is an appropriate methodology when utilized correctly.

Hey all,
Another news item relevant to this thread:
http://www.guardian.co.uk/commentisfr...

It's a Guardian article that concurs with William Udall's warning that neuroimaging can't locate things like jealousy or fundamentalism.

sorry: it's spelled Utall

I guess I should qualify my view on EMERGENCE. I was using it in the greater context of the Mind which exists at birth and not to be construed as including intellect. In terms of advanced functioning, as in coping with the external environment and survival, I have no problem with the concept of Emergence. I have no problem because of the associations that get established between experiences subsequently after the fact - it's what gives the moment meaning. Whether or not Emergence is dogma depends on the context in which it is used.

When it comes to FREE WILL, the jury is still out for me. Society and its body of laws go to great lengths to control and limit Free Will. By virtue of the inhibitory networks, we are able to exercise some control over impulse (most of us). One could make the case either way for the existence of Free Will since there are predisposition for things like altruism, empathy, sympathy, values, and mores (which will always have some degree of influence even in the antisocial personality). There's such a cascade of influences, it's hard to say if it exists. I have to ask myself, what is the difference between a nearly infinite number of ways to respond and Free Will - are they the same? There are many naturally occurring constraints that would limit what's possible, both internal and external, so your guess is as good as mine.

As for "hauling down the reductionist flag," I was mainly referring to that in the context of "Complex Systems." If you want to understand the system, you have to focus on the system level. Reductionism is needed to characterize the elements that make up the system and realize their limitations. It helps to formulate theory about how the system interacts as a whole by whittling down what's possible. Unfortunately, reductionism tends to become a way of life for researchers; they tend to focus on the minutia out of context with the whole. They split hairs for the sake of splitting hairs - see theories concerning microtubules as wave guides for intercellular communications (Yikes).

Let me revisit diffusion tensor tomography. Remember, you are looking at a macroscopic view of something operating on a microscopic level. The image is 2D while the brain is 3D, so you are only looking at a single plane. The surgeon can look at multiple planes to get an idea about the area he wants to treat and that gets him into the neighborhood, but which house? Well, it's useful for him, but for research purposes? I really believe (for the time being at least) that the most important tool in the neurologists kit is "Imagination," however, that can lead them astray if their perspective is wrong and there are a lot of different perspectives that one can have. The pet theories we have invested in tend to skew our perception into vested positions – like blinders on a horse, they narrow the focus.

Well, I’m glad to see that I’m not the only one to see and express the existence of dogma in science.