What Does Any Part of the Brain Do?

By Neuroskeptic | March 9, 2018 2:43 pm

How can we know the function of a region of the brain? Have we been approaching the problem in the wrong way? An interesting new paper from German neuroscientists Sarah Genon and colleagues explores these questions.

dreams

According to Genon et al., neuroscientists have generally approached the brain from the standpoint of behavior. We ask: what is the neural basis of this behavioral or psychological function?

Traditionally, assigning functions to brain regions has mainly been based on conceptualizations of functions from many different disciplines that are interested in the study of the mind and behavior… ‘speech production’, ‘perspective taking’, and ‘emotional regulation’ are a few examples of these behavioral functions.

This “behavior-first” approach has revealed many associations between particular functions and particular brain regions. However, Genon et al. say, it has become clear that any given behavioral function involves more than one brain region, and it may be that there is no ‘necessary and sufficient brain area’ for any behavioral function.



So, the authors say, we may need to adopt a “brain region-first” perspective. Instead of looking for the neural basis of a behavior, we should instead be asking: what does this brain region do? What functions is it involved in?

For any brain region, we can think of many different behavioral functions, based on the perspective from which we consider this brain region. In practice, most of these behavioral functions can somehow be related to one another and seem to comprise a core computational function (i.e., an operation-function) that grounds all behavioral associations but remains latent and is not directly observed.

In other words, any brain region will be involved in many functions, but by triangulating between them we can (hopefully) arrive at the true core function of the region:

Our current knowledge of the functional specialization of a given brain region can be conceptualized as a polyhedron with its many sides (i.e., many behavioral functions), the sum of which can only be appreciated by investigation from many different perspectives, but whose core center remains intangible.

functional_polyhedron

Genon et al. say that tools like NeuroSynth are ideal for helping us to map out these polyhedra. NeuroSynth is a database of published fMRI studies and it allows us to adopt a “brain region first” analysis: we can select a brain region and see what kinds of tasks and stimuli have been known to activate that region (although with many caveats). BrainMap is a similar tool.

This is a fascinating paper. I wonder, however, whether we might end up discovering that all brain regions – or at least, the bulk of the cerebral cortex – have the same core cognitive function? It might be that most of the cortical ‘modules’ are actually doing the same kind of processing, but operating on different inputs.

To give a simple example, a brain area which mainly gets auditory input will light up in response to different stimuli compared to one that gets visual input, but they might be doing the same basic operations on the input. I don’t think anyone knows, yet, whether different cortical areas are truly doing different computations.

CATEGORIZED UNDER: fMRI, methods, select, Top Posts
ADVERTISEMENT
  • Cyril

    i got your point but different areas have different architectures (cyto and myelo) which leads to believe in different computations. The cerebellum is good counter example, different regions are involved in different processes yet the same cytoarchitectony suggests same computations – back to the paper, it’s not a completely new idea…

    • http://blogs.discovermagazine.com/neuroskeptic/ Neuroskeptic

      Yes, the cerebellum is probably doing something different, but cortical regions (except V1 (?) ) are pretty similar, I think? Only subtle cyto differences…

      • http://www.pbase.com/davidjl David Littleboy

        There’s an ongoing argument here, with Jeff Hawkins arguing that the similarities between cortical columns mean they’re all doing essentially the same thing and Gary Marcus arguing that the differences mean they’re doing substantially different things. For the nonce, I found Marcus’ arguments more persuasive, but it’s nice to see people talking about actual neurons and actual structures actually seen in the brain.

  • http://jamesmacshine.com Mac Shine

    I love the idea of a core computation that varies according to inputs, but why limit it to cortex? I imagine that the proportional influence of the subcortical regions that connect to each cortical location (particularly from thalamus, striatum + cerebellum) play a huge role in determining the information processing capability of different cortical regions.

  • johnLK

    The function of a brain region is its input/output function. in other words, what algorithms does it perform with what information. The linkage to behavior is indirect, and clearly has to go thru many regions to reach an output. We’re a long way from understanding this in any brain region. Linkage to behavior may be a key to understanding, but its not “the function”.

  • CL

    https://uploads.disquscdn.com/images/324124e336d885891490496b21651d19901c6f6456a49f74efa272a2ec550338.jpg

    I like the general approach, but consider this single pyramidal neuron in the mouse cortex (eLife, Economy 2016, https://elifesciences.org/articles/10566 ).
    Regional explanations will not get us very far, as we are dealing with networks in any more complex situation than perhaps primary sensory processing.

  • My Own Life

    The functional polyhedron of the hippocampus illustrates a recurring deficiency in human brain studies: why is its role in our emotions ignored? Animal studies have no problem assessing that specialized hippocampal areas regulate emotional states.

  • http://www.mazepath.com/uncleal/EquivPrinFail.pdf Uncle Al

    Brain function is emergent. Sufficiently close examination disappears it.

  • FSE

    This is just a restatement of the old “Reverse Inference Fallacy”. The authors propose to get around it by building and testing more refined conceptual models, but that misses the point. It’s not fundamentally a problem of model accuracy, it’s fundamentally a problem of epistemology. Even if you can show that every experimental activation of the hippocampus can conceptually be linked to X, that does *not* mean that an unexplained activation of the hippocampus must also be linked to to X. Every time you see me at work, I am wearing pants, but if you see someone at work wearing pants, you can’t necessarily conclude it’s me.

    If you’re interested in the details, this is a great paper:

    https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3240863/

    • typically

      Wearing pants is a fairly ubiquitous function, liking breathing. It is indeed difficult to infer on a function X that is linked to activation everywhere in the brain. However, if when X is performed we see increased activation only in hippocampus, this is evidence that the two are linked. If every time you see me at work, I am doing spreadsheets, it is a safe bet to conclude I work in Accounting.

      • FSE

        The point is that you need to be very careful when you state your inference, and often in neuroscience people are careless.

        If every time you see me at work, I am doing spreadsheets, then you can infer that the next time you see me at work, I will be doing spreadsheets. You should not infer that I am doing spreadsheets when I am not at work. Nor can you infer that I am the only one doing spreadsheets.

        It is possible that I primarily responsible for something you have not found a way to observe. For instance, perhaps I spend most of my time taking care of my children, and I come into work every so often to help out the accountant.

        In the brain, it is valid to infer things like “memory tasks activate the hippocampus.” But it is not valid to make reverse inference, “when the hippocampus is active, the brain is engaged in a memory task.” It’s quite possible that the hippocampus is doing something totally unrelated, that we are unable to reliably elicit by experiment.

        • typically

          Very much agree with all of that.

  • Pingback: Specific parts of the brain do many things | Psychology for Designers()

  • Pingback: What Does Any Part of the Brain Do? « Adafruit Industries – Makers, hackers, artists, designers and engineers!()

  • Pingback: What Does Any Part of the Brain Do? - HudsonWerks()

  • Pingback: AKETXE Consulting – What Does Any Part of the Brain Do?()

  • Pingback: JOIN&WIN – What Does Any Part of the Brain Do?()

  • Telos

    This is not unlike what Alexander Luria was postulating over half a decade ago in the Soviet Union

  • Pingback: PT1: What really is Cognitive Type? | Cognitive Typology()

NEW ON DISCOVER
OPEN
CITIZEN SCIENCE
ADVERTISEMENT

Neuroskeptic

No brain. No gain.

About Neuroskeptic

Neuroskeptic is a British neuroscientist who takes a skeptical look at his own field, and beyond. His blog offers a look at the latest developments in neuroscience, psychiatry and psychology through a critical lens.

ADVERTISEMENT

See More

@Neuro_Skeptic on Twitter

ADVERTISEMENT

Discover's Newsletter

Sign up to get the latest science news delivered weekly right to your inbox!

Collapse bottom bar
+