I once gave a (perfectly awful) cognitive science lecture at a major centre for brain imaging research. The main project there, as best I could tell, was to provide subjects with some or other experimental tasks to do and take pictures of their brains while they did them. The lecture was followed by the usual mildly boozy dinner, over which professional inhibitions relaxed a bit. I kept asking, as politely as I could manage, how the neuroscientists decided which experimental tasks it would be interesting to make brain maps for. I kept getting the impression that they didn’t much care. Their idea was apparently that experimental data are, ipso facto, a good thing; and that experimental data about when and where the brain lights up are, ipso facto, a better thing than most. I guess I must have been unsubtle in pressing my question because, at a pause in the conversation, one of my hosts rounded on me. ‘You think we’re wasting our time, don’t you?’ he asked. I admit, I didn’t know quite what to say. I’ve been wondering about it ever since.
See also:
Grill-Spector, K., & Weiner, K. S. (2014). The functional architecture of the ventral temporal cortex and its role in categorization. Nature Reviews Neuroscience, 15(8), 536-548.
ABSTRACT: Visual categorization is thought to occur in the human ventral temporal cortex (VTC), but how this categorization is achieved is still largely unknown. In this Review, we consider the computations and representations that are necessary for categorization and examine how the microanatomical and macroanatomical layout of the VTC might optimize them to achieve rapid and flexible visual categorization. We propose that efficient categorization is achieved by organizing representations in a nested spatial hierarchy in the VTC. This spatial hierarchy serves as a neural infrastructure for the representational hierarchy of visual information in the VTC and thereby enables flexible access to category information at several levels of abstraction.
Fodor makes an excellent point in his article: what will change in our understanding of the ‘mind,’ of ourselves, or of the world, if we are able to localize ‘objects’ in the brain? He reveals the impact on trends in science, obsessively rooted in positivism: why publish what we *don’t* know when we can share what we *do* (even if what we do know has no impact on our field of study or on the world)? It is easy to forget to ask ourselves the ‘why’ of what we are doing. In scientific (empirical) observation and study, forgetting to ask comes at a huge cost, often literally. And in studying what we already know to be true (for example, that nouns are different than verbs) we seem to be avoiding what fuels curiosity-driven science: not knowing the answer.
ReplyDeleteNeuroscience, perhaps among the trendiest fields of scientific study today, is alluring both in research and in the public domain. It is extremely historically fresh, and there is very little we can admit to knowing on the topic. And what we do ‘know’ is constantly challenged and overturned, multiple times in the span of one person’s life.
Publishing what we do ‘know’ says less about neuroscience and more about us: we would rather be secure in our knowledge (despite its helpfulness) than not know to begin with. Science and ego are not mutually exclusive.
Fodor's main point is: You cannot reverse-engineer "how and why" from "when and where."
DeleteFodors point that science should be sparing with public money is really valuable -- the natural attractiveness of studies that show us what parts of the brain light up when definitely deserves scrutiny; it’s obvious that if neuroscience failed to come up with anatomically distinct substrates for the concepts of trains and buses, we wouldn’t question that those categories are separate.
ReplyDeleteEven given the power of Fodor's argument, I have a hard time believing that establishing a functional architecture of the brain (or pretty much any system) can’t do wonders for our understanding of the system and for our ability to reverse engineer it. I’d argue knowing the structures that a carburetor (or a heart, or the primary visual cortex) is connected to, and what those structures are connected to (where different tasks are managed in relation to each other) is vital to a working understanding of the system, regardless of whether you intend to take the part out. In the case of the visual system, research has yielded interesting and actionable truths about how the brain processes information. Researchers have observed a connected line of anatomically distinct areas that represent visual information at different levels of abstraction where neurons are shown to respond selectively textures, edges, features, and even vastly different images of people. Of course this is far from the whole story on what is really just a small component of what the brain can do, but how would information about certain structures function and connectivity not bring us closer to understanding/reverse engineering the things the brain does?
I agree with your comment Kevin. I also have a hard time believing Fodor's argument that nothing can come out of these studies. I'm wondering for the auditory track as well, were doctors not able to reverse engineer cochlear implants by learning how the auditory system works? This could potentially be done with other systems as well, showing the usefulness of this research.
DeleteFodor agrees that neuroimaging is valuable for clinical purposes; what he questions is whether it can help solve the easy problem -- how and why does the brain produce this cognitive capacity? -- to find out when and where it does it.
DeleteThe challenge in causal modelling of the brain's cognitive capacity is whether cognitive neuroscience can provide a way to generate a cognitive capacity that enables modellers otherwise did not know how to generate.
The cochlea is not perform a cognitive function but a vegetative one, like the heart.
Even if cognitive neuroscience research doesn't present the whole picture on solving the easy problem, I still feel like information around the divisions of the brain, their functional specialization & where they connect will prove helpful. In solving any large problem, its essential to divide the solution between meaningful and non-arbitrarily connected subsystems -- In solving some difficult problem, access to the general framework a master solver used to successfully complete the task, how their solution divided labor etc., will give you a leg up. Even if you were asked to write a computer program simulating a game of cards --an unbelievably simple problem-- It would noticeably cut your solving time to hear about the subsystems (to compare hands, keep track of bets/money/turns, generate random cards, etc) a successful solution employed and how they were organized hierarchically.
DeleteThis rests on the assumption that general info about the functional specialization and organization of a well defined computational system would be analogous to that of the brain, so that’s probably something I should hear more about.
Yes, the way the brain modularizes components of our cognitive capacity might help reverse-engineer the causal mechanism: Can you give examples of where it has actually done so?
DeleteI also agree with your comment, Kevin. Bear with me as I know nothing about neuroscience or car engines, but Fodor’s comparison to the car engine – that knowing where the carburetor is located is more or less arbitrary in understanding how the engine works – seems to be a weak comparison to the potential usefulness of brain imaging to me. We do not need to reverse engineer the car engine, because we designed it and can explain it, more efficiently and effectively than ‘re-inventing’ it. If a car engine were a naturally occurring phenomenon that we wanted to understand by reverse-engineering, maybe a good starting point would be to make a map and discern which parts are connected to / interact with which other pieces – and which parts of the engine are active when the car is performing different functions.
Delete...if cognitive capacity were like vegetative capacity (e.g., the heart, or the lungs or temperature regulation), where structure is a clue to function. But is that true for cognitive capacity, even in as "simple" a case as mirror-capacity and imitation?
DeleteThe concept of Convolutional Neural Networks, the computational framework that underlies highly accurate image classification & facial recognition, seems to have been modeled after or at least inspired by neuroscience research into the organization of the visual system (Hubel and Wiesel etc.). That these systems have individual units that respond selectively to certain features at different levels of abstraction strongly suggests that their development owes something to this neuro research right?
DeleteProfessor, to your point that the vegetative function of an organ such as the heart or the lungs is evident from its structure, perhaps our imaging techniques are simply not yet sophisticated enough to reveal the similarly evident function of the brain? Consider the solar system, for years the very best scientists believed that the Earth was at the centre of the solar system, as our imaging and tools became more precise and sophisticated, slowly our understanding of the universe grew with it.
DeleteWhile we may believe that our understanding of the physicality of the brain to be quite sophisticated today while still not having a direct link to cognition, perhaps in twenty years Fodor will look the fool when something previously unseen reveals itself in the physical structure.
Hi Kyle! I was also thinking about this topic since the lecture. I can see the reasoning behind how the improvements in astrophysics showed us that our ideas about solar system was wrong; and so the future improvements in our imaging techniques might provide us with better answers about how the brain works.
DeleteBUT, when we try to answer how the brain does what it does, I think the important question we are dealing with is not directly about structure, and functional processes of the brain. We also still need to answer; can we clearly figure out where the “feeling” originates from given this new knowledge that we obtained through better/finer imaging techniques.
Thinking about this argument about better imaging techniques, it seems like, the “feeling” should be buried somewhere in the structure and functional part of the brain. From what we have discussed in class until this point, (currently) I think that it is an implausible idea. Because, even if we do completely understand the brain through these techniques, these techniques alone will not provide us with enough knowledge to be able to “decide” whether we found how the “feelings” originate in the brain (unless these techniques completely change the way we think about how brain-as an object- & mind interacts)
“What part of how your engine works have you failed to understand if you don’t know that?”
ReplyDeleteThe argument Fodor makes is interesting. Is there a point to trying to figure out where in the brain, specific processes occur if we cannot extrapolate how these thoughts occur from the where. Although I somewhat agree with Fodor when he discusses the post-lecture dinner party and the seemingly frivolous nature of repetitive testing with subjects in an arbitrary manner, I disagree that the where can never explain the ‘how.’ Particularly in cases where lesions or brain damage has occurred, the ability to know where a lesion has occurred and what function that brain area is responsible for can be instrumental in explaining the ‘how’ and especially the ‘why.’ In Ramachandran’s book, “Phantoms in the Brain,” he explores a variety of disorders that arise from brain lesions and disorders and attempts to help ease some of the mystery surrounding these issues as well as occasionally ease some of the discomforts of the people experiencing the symptoms. Most of the time, he only manages to do this by performing these repetitive, albeit not arbitrary tests, to determine the function of the injured brain part. If even only for the improved conditions of these patients, the study of where something occurs is still necessary.
Fodor does not question clinical usefulness.
DeleteIn this article, Fodor questioned the reason why everyone is so enthusiastic about neural imaging considering it's expensive -- as science is all expensive -- and the resource is quite limited. "But whereas, historically, studies of the localisation of brain functions have often been clinically motivated, I take it to be currently the consensus that they have significant scientific import over and above their implications for medical practice." Here while he admits the clinical motivation of neuroimaging studies, he is actually directing his arguments towards their usefulness "over and above medical practice". He also admits that it's meaningful to find "whether mental functions are neurally localised in the brain" since it's indicative of how the mind works, whereas cannot be convinced that it's also significant to study "where they are neurally localised in the brain", above medical usage of course.
ReplyDeleteFurthermore, he points out that "in fact there’s no good reason to think that similarity of psychological functions generally predicts similarity of brain locations or vice versa.", which seems to be trying to say that even if we accurately matched some of the different neural counterparts with their distinct, corresponding mental states, chances are that we still would not be able to get any luck in figuring out how on earth the brain manages to trigger these mental states and more importantly, generalizing these findings to explain other mental states. To conclude, neuroimaging does not seem like a promising investment within the arena of cognitive science.
Short version of Fodor's point: "Where and when something happens in the brain does not help explain how and why the brain makes us able to do what we can can do." Is that true? And if so, is it just temporarily true, till we find out more? Or is functional localization in the brain not helpful in reverse-engineering the causal basis of the brain's cognitive capacities (T3)?
DeleteThe questions posed by Professor Harnad reminds me of issues discussed by the mind-body dualism problem. Fodor questions "whether mental functions are neurally localised in the brain" -- to which Descartes would answer that "mind and soul of the man [aka mental functions] are entirely different from the body [or brain]." Perhaps Fodor did not intend to venture into philosophy, as he admits that "dualistic metaphysics was now out of fashion" in neurology; however, I still wonder if he is making an argument for dualism by challenging the intellectual importance of localized brain functions.
DeleteTherefore, to answer Professor Harnad's question: I believe that we may not be able to answer whether where and when something happens in the brain helps us understand the how and why until the hard problem is solved. The hard problem of cognitive science, understanding the how and why of what we feel, essentially mirrors Fodor's main point and once solved, also answers the mind-body problem. However, there is a reason why this is called the hard problem; we cannot even describe what "feeling" feels like, much less generalize the experience in a way that one can study it. Functional localization is a stepping stone that may not necessarily get us to the hard problem but technologies like brain imaging and studies in neuropsychology gets the wheel turning. We simply cannot resort to sitting in our armchairs. Reverse-engineering at least attempts to answer the easy problem…and may drop clues leading to the hard problem.
“But given that it matters to both sides whether, by and large, mental functions have characteristic places in the brain, why should it matter to either side where the places are?”
ReplyDeleteKnowing where mental functions are located in the brain is useless because this knowledge doesn’t us help explain how our cognitive capacities are generated (the ‘how it is that we can do what we can do’). Fodor doesn’t deny the importance of neuroimaging techniques in clinical and medical practice. It is crucial for a neurosurgeon to know where an individual’s Broca’s area is located in order to avoid inflicting irreparable damage to that person’s ability to produce language. But does this knowledge and the discovery that when we think about teapots an area above our left ear lights up tell us anything about cognition? The answer is no, they don’t, since these facts are merely informative and fail to provide us with a causal mechanism/explanation underlying cognitive processes.
“But why (unless you’re thinking of having it taken out) does it matter where in the engine the carburettor is? What part of how your engine works have you failed to understand if you don’t know that?”
It doesn’t matter where the carburettor is located, as much as it doesn’t matter where MNs are located (or whether the genetic or associative account is more accurate in explaining how and why MNs come to be present in the brain). As elaborated above, we need to come up with a causal explanation for cognitive processes to be able to answer Cognitive Science’s easy question.
So if not by studying the brain, how to reverse-engineer cognition?
DeleteAs mentioned in some of the previous skywritings Fodor calls into question the importance of determining brain geography in the study of the mind. He concludes that clinically, knowing what different sections of the brain do and where they are located is very important however there is no use when studying the mind. This makes sense, because if cognition is trying to figure out what goes on between A & B, knowing where A & B are in respect to one another isn't very useful.
ReplyDeleteIt was also interesting how Fodor introduced the debate between empiricism and rationalism. Empiricists, as they won the phrenology battle, seem to be in the lead when it comes to studying the brain. To me, their theories of overall functional equipotentiality seems to line up with the ideas of computationalism. This is because according to empiricists, no matter where the “all mental processes reduce to patterns of associations”. It seems plausible that these “patterns of associations” are synonymous to computations if they are simple enough. A question then I have for Fodor is why he thinks that empiricists won’t win the battle? Could it be because of Searle’s critiques on Strong AI?
The empiricist/rationalist battle is as irrelevant here as it is in the mirror neuron article by Cook et al (4a). The objective is to reverse-engineer cognitive capacity, not to fight or win abstract battles. "Association" is trivial, and is not a causal mechanism, both for explaining what "mirror neurons" do to give us mirroring and imitation capacity and for explaining anything else in our cognitive capacity other than how paired-association of nonsense syllables might work.
DeleteFodor was a computationalist (in his co-authored articles with Pylyshyn), so, yes, he is refuted by Searle's CRA. Yes, it might be possible that in some vague way he equated association with computation...
“But whereas, historically, studies of the localisation of brain functions have often been clinically motivated, I take it to be currently the consensus that they have significant scientific import over and above their implications for medical practice.”
ReplyDeleteWhile I can certainly understand Fodor’s impression of localization studies being frivolous, I do not agree with his judgment that they are a waste of research funds. In the above quote, he preemptively mentions how the consensus is that these studies have more than just clinical implications. While that may be the case, it seems like he’s arguing that localization studies have no real purpose while openly admitting and then dismissing one of its main purposes. I think its clinical applications is reason enough to conclude that localization studies are not a waste of research funds.
In addition, Fodor mentions a fable whereby Pavlov tells a little girl that he’s drilling holes in the mouths of dogs to show that they will salivate when expecting food. She states that we already know that, and he responds with “Now we know it scientifically”. Fodor then maintains that the same kind of reasoning is behind the motivation of brain localization research. However, it’s quite clear that showing that dogs salivate was a means to an end for Pavlov, which was to uncover the underlying mechanism that causes dogs to salivate in response to food, so that he can further explore that mechanism. This is similar to brain localization, just knowing the location of a certain mental function is pretty useless but taken together with other research may reveal something more. For instance, knowing that emotion processing occurs in the amygdala doesn’t tell us much alone, but if we later see that people with generalized anxiety disorder have a heightened amygdala response to threatening stimuli, then we learn more about that disorder. Or if someone has a damaged amygdala, we can reasonably expect them to present with symptoms of impaired emotion processing.
I do agree with Fodor that localization studies may, at times, appear to serve no practical purpose, but I think it was prematurely dismissive of him to assert their fruitlessness, especially considering that if new technologies develop, then new opportunities with old information arises. A perfect example of that would be the growing field of artificial intelligence and its use of human cognition as a template. Also, I’m not even certain that any researcher would investigate and publish something as trivial as which part of the brain lights up when thinking about teapots specifically.
Fodor's critique was about localization in cogsci, not clinical research. Not sure how you think AI can make localization help...
DeleteI understand that now, in fact, if he isn't including clinical research then I would agree with him that localization would not be of much use, similarly to how EEG isn't really informative with regards to any specific form of psychopathology (other than vulnerability or risk factors). If we're talking about localization alone, then yes I agree that it cannot be of much use.
DeleteHowever, I do believe that there may possibly be some potential use for localization to help advance the field of AI, if it were to inform us, for instance, that certain functions that we thought were one, are actually two different functions or aspects that come together in the brain. For example, Broca's area and Wernicke's area are two different areas in the brain associated with language, we were able to disentangle speech production and speech comprehension. Thanks to this, we are able to know that understanding the meaning of language is a different but closely related aspect from producing language, and that they eventually come together, but are different actions in the brain. This could serve as a sort of reverse engineered template of how cognition works, for instance, speech production (motor action) draws on the comprehension of language in order to produce that specific action. Moreover, understanding prosody, which would be essential for an AI to be able to do if it is to communicate indistinguishably as a human, has been shown to be located not only on a different location of the brain than language comprehension, but also in a different hemisphere, so that the literal and emotional aspects of language are two different things that come together before ultimately understanding speech. This could help researchers in the field of AI to serve as a framework for language comprehension, whereby emotional understanding would be necessary for an AI, as well as semantic and syntactic understanding, and that these things should be understood separately before finally coming together to give rise to language comprehension tantamount to the language comprehension that most humans have.
Moreover, the discovery that visual perception in the brain arises out of a hierarchy of feature detection could also serve as a reverse engineered template for perception in the field of AI and this is directly related to brain localization, whereby simpler features of visual input are detected in slightly different areas of the brain than more complex features in a hierarchical manner.
Re: “whereby simpler features of visual input are detected in slightly different areas of the brain than more complex features in a hierarchical manner.”
DeleteI think you raise and interesting point that (to me at least) somewhat echos Harnad’s (2001) argument. Towards the end of the article, Harnad lists 3 points on which Searle overreached, one of them being the conclusion that Searle draws that cognition is not computation at all. Harnad argues that Searle has only proven that that computation cannot explain everything about cognition, but it still explains part of the problem.
Going back to your comment, I think many mental process have a higher-order aspect to them which require interpretation and understanding, but most of them, especially within the context of sensory perception, also have a lower-order and more basic aspect. Breaking them down into smaller parts based on the localization of function can help tease out the separation and the point at which computation no longer suffices, which could help support or dismiss Harnad’s (2001) point about computation accounting for some of the process.
Neuroscientists have attributed specific brain regions to specific functions: they have located certain regions of the brain that are consistently shown to be responsible for the ability to perform certain actions or receive certain inputs. (For example, if there is damage to a person's primary visual cortex, they will be unable to see.)
ReplyDeleteI agree with Fodor when he says that it is of little import to understand *where* a brain function occurs. In fact, a successful model of cognition (a T3 device) would be successful if it's visual input were acquired on the bottom of its left foot, as long as it got it somehow: as long as the machine can do all the things that we can, it is a successful model of cognition.
Researching where functions are localized is an unimportant task. What matters is only that a system functions the way it does.
And how do we find that out?
Delete“But why (unless you’re thinking of having it taken out) does it matter where in the engine the carburettor is? What part of how your engine works have you failed to understand if you don’t know that?”
ReplyDeleteFodor’s point can be roughly resumed as “Knowing where and when the brain is active during a cognitive process does not explain how and why this cognitive process happen.” Set aside clinical purposes (such as not impairing capacity X while removing a brain tumor, let’s say), the study of the brain should not be the main way for investigation in cognitive science.
This point of view is closely related to the second point of computationalism (brain states are just implementation-independent implementations of computational states, i.e. the brain is irrelevant). But if you accept the Chinese room argument, you accept that full computationalism cannot work. So what’s left? Is the structure of the brain relevant or not? We could argue that the structure of the brain (and the nervous system in general) is (at least) relevant for some analog processes (e.g. the structure of the ear can be described like a causal mechanism explaining sound hearing, if we let aside the explanation of the “feeling” of hearing).
Nevertheless, I would say that Fodor’s point is relevant when we try explaining more abstract activities like memory, attention, executive functions, etc. That returns us to T3 being the main way to verify our explanations. But that makes me wonder what we are trying to explain. We could speculate that it is our very conception of “what a human can do” is ill-conceived. That notions like “speak”, “move” are too vague to be useful for T3, and that a more precise (whatever that means) account of our capacities would be beneficial. But I think that we cannot really do better right now, because the “what can we do” is precisely what we are trying to understand.
The trouble with studying the brain is that, unlike the heart, its structure does not give a clue to its function. It is very likely that some T4 functions will be needed for T3, yet it is not clear how localizing function in the brain reveals what the locus is actually doing, and how.
DeleteI am very inclined to agree with many of Fodor’s points. Although I do not discredit the brain in being part of human cognition, I feel that it is over-sufficient. If we were to reverse-engineer the brain as a goal of understanding human cognition, sure, we might find that motor perception/production is traced back to the frontal lobe, somatosensory processing is traced back to the parietal lobe, auditory processing at the temporal lobe, and visual processing at the occipital lobe. On top of this, there are also findings that the amygdala is in part responsible for emotional processing, the hypothalamus biologically regulates us, the hippocampus “forms” memories, and the pituitary gland is the hormone dispenser. These are some of the most common findings that are reported to be evident in every typically-developing individual. But there are some “processes” that are still poorly understood in terms of where and when they happen in the brain. Even for higher-level emotional processing, it is hard to trace where in the brain that happens. It varies between individuals and every group of researchers that write about it seem to name a plethora of areas (I won’t name any of them because there are some that I didn’t even know existed!).
ReplyDeleteSo, if we cannot trace exactly where, for example, higher-level emotional processing occurs in the brain, is it really that important? If it is so variable among individuals, can’t we just say “we know it happens but we will never know where in the brain it happens”? There would be so many felt states that are, not only unique to individuals, but also incredibly variable. I might have had felt states that you have never felt and may never feel. So, if this hard question of cognition (i.e., how we feel when we do things) is pretty much unsolvable, shouldn’t we only be considered with basic processes (hows and whys) instead of reverse-engineering absolutely everything a human brain can do?
To localize function is not to reverse-engineer it -- except for vegetative (rather than cognitive) function, where the structure resembles the function.
DeleteIn the article, Fodor questions the usefulness of investigating neural localizations of mental function. While he concedes that these types of studies are useful for clinical purposes, he claims that it doesn’t matter where the mental function occurs; he is only interested in the how and why. One of the replies to Fodor is as follows, “A theory which integrates brain and mind processes will be a major goal for neuroscientists, psychologists and philosophers in the coming decades. It will need to understand both the particularities of the micromechanisms of nerve cells and their interactions and the dynamics of the system as a whole”. I agree more with this stance than Fodor’s argument. If the goal of cognitive science is to reverse-engineer cognition, the best place to study is the brain. The where and when questions help answer the how and why questions of cognition science. For example, investigating disparities between neural connectivity and localization is important when studying psychiatric disorders. Why do certain people with schizophrenia experience hallucinations while others only experience paranoia? Identifying differences in brain anatomy and physiology is insightful when considering how these disorders manifest. Studying functional localizations and connectivity in the brain could potentially illuminate not only the difference in experience between the control individuals and those with the disorder, but also the range of experiences that those with the disorder can have.
ReplyDeleteFodor agrees that localizing function is useful for clinical purposes. But can you think of examples of how it has helped explain the causal mechanism of (normal) cognitive function?
Delete“Brain scientists are supposed to be materialists, and materialists are supposed not to doubt that distinct mental states have ipso facto got different neural counterparts. That being so, why does it matter where in the brain their different counterparts are?”
ReplyDeleteI think that the attempt to spatially localize mental functions has impeded neuroimaging research. Neuroimaging is a powerful tool to investigate cerebral activity. It has no doubt a clinical potential when used as a tool, not as glasses to read the mind.
If we consider pain research, there is a controversy about whether there exists brain tissues specialized in pain processing. Using fMRI, researchers have highlighted several brain regions that are active during painful stimulation. This has led to the creation of a “pain matrix”. But these areas have been found to be equally activated by salient stimuli of other modalities, weakening this holly concept of “pain matrix”. Still, this “pain matrix” theory has led to some erroneous/unsound conclusions that social cognition might result in pain because empathizing to someone’ else pain has yielded activity in these same exact (non pain specific) regions.
Beside spatial investigation, neuroimaging can be used as a tool to investigated temporal rhythms. If scientists, as devoted materialists, agree that mental states are embodied, wherever that embodiment is it must come with a distinct timing of neural firing. Neuroimaging can be used to investigate brain rhythms and causality between stimuli and evoked potentials/spectral power recorded all over the brain.
I think that there is a difference between attempting to find a “teapot” locus using fMRI and using electroencephalography (EEG) to assess synchrony and coherence between widespread neural population.
PS: I am doing a project investigating biomarkers of acute pain using EEG so I am optimistically biased toward the clinical utility of neuroimaging, *when used as a tool and not as a mean to read the mind.
Pain is mostly a clinical problem, so neuro-imaging has a good chance of being helpful.
DeleteBut what makes you think timing (i.e., when) is any more likely than localization (where) to reveal how or why (for non-vegetative function)?
I understand that Fodor’s point is that we cannot reverse-engineer the ‘how’ and the ‘why’ from neuroimaging data that only allows us to observe a ‘when’ and a ‘where’.
DeleteMy opinion is that in a weak AI minding, if we were to use neuroimaging data and machine learning as a tool to understand how the brain works by means of simulation, I think we would reach a better accuracy using the ‘when’ than the ‘where’. Particularly, if clinicians and neuroscientists are looking to infer (I say infer because I acknowledge that right now the technology doesn’t allow us to know at 100%) causality from activity in different brain structures, the timing is crucial.
We know that there is electrical activity in the brain that is modulated by internal and external stressors. Even Fodor agrees for the sake of his arguments that consciousness is correlated with brain rhythms at around 40 Hz. I think the problem is to try to compartmentalize this activity into pre-established anatomical landmarks. Instead we should look into synchrony and coherence between these different brain rhythms that are modulated by stimuli we can observe by means of sensorimotor interactions with the physical world. Electrical activity is something we can objectively and directly measure. Stimulus onset is also something we can objectively and directly measure. I don’t understand the need to link this activity to special loci of the brain we can’t directly observe, considering fMRI operates on level of deoxygenation in the blood. I think inferences should be based on what we can directly and objectively measure and observe. And for that sake, I think investigating the ‘when’ has more potential that the ‘where’ to help us understand how the brain works.
I must admit, I have quite the gripe with this article. Potentially, it's the bias that comes with pursuing a career very obviously focused on fMRI and variations on brain mapping. I think the picture painted here of neuroscientists and experimental psychologists looking at brains is unrepresentative of the actual work that goes on, or at least is representative of a very narrow scope from the author. The last paragraph of his text describes a specific kind of research. I'd argue that a very small minority of people looking into the brain actually think this way (where regions of interest are randomly selected). The reality is that a lot of the brain research that was ongoing and that would have been published around the time of this piece (1999) has been seminal in leading research that has had huge consequences on populations like psychopathological patients. We need to map out brain regions to be able to use fMRI tools (one of the only methodologies that doesn't rely on participant ratings) to assess behavioral data. If I'm looking into differential access of autobiographical memories in patients with depression, I need to know where all of the components of autobiographical memories are "located" to make sure that my participants are actually accessing the same memory information from different processes. That would be scientifically impossible without the basic mapping research that Fodor seems to hate so much.
ReplyDeleteI'd also add that there is merit to knowing something "scientifically". The author describes the horrific research techniques employed by Pavlov to make this point, but the Pavlovian reply in this context is fair in my opinion. The little girl in his story (who asks Pavlov why he's drilling holes in the mouths of dogs) says that we already knew dogs salivate. That "knowledge" is intuitive, yes, but is also purely circumstantial. The example of salivation in the presence of food here is too simple. A better example might be something like the fact that people who suffer from depression tend to be biased towards negative information. They respond quicker to it than positive or neutral stimuli. That might seem slightly intuitive: we can make hypotheses about why this would happen and it generally makes sense. The reality, though, is that treatments that involve this bias are not going to become widespread unless we prove "scientifically" that the bias exists in the first place (i.e. we are certain that the found bias could not be random, it definitely is because of a patient’s depressive symptoms). If we based psychiatric and psychological treatments on intuition, we'd be wasting a hell of a lot of resources on those treatments, and in Fodor's ideal world of psychology research, we'd have an entirely different problem with resource allocation.
All in all, Fodor’s ideas are relevant, and I like his comparison with the rationalist VS empiricist debate, but he very clearly misses the point of all of the brain mapping research if he thinks its only application is stuff like brain surgery.
But Fodor says from the outset that he is not questioning the usefulness of neuro-imaging in clinical research; just its usefulness in reverse engineering ("neurotypical") basic cognitive capacity (the "easy problem" of cognitive science).
DeleteThe existence of classical conditioning could have been discovered and described with far more merciful experiments than Pavlov's, non-invasively (and in humans rather than experimental animals). His use of dogs was more indicative of the instrumental attitude toward animals as scientific materials rather than feeling beings. (And neither Pavlovian associationism nor Skinnerian reinforcement learning has taken us very far insofar as the "easy problem" is concerned, despite the cost in animal suffering.)
Fodor was a philosopher, and a very clever curmudgeon, but not a very constructive one. His scoffing at neuroscience enraged many people. Yet he had a heart. I will never forget when he told me, obviously grieving after his cat died, "I never thought you could care so much about a cat." That was not that long before his own death. He also loved opera (and sailing with Zenon Pylyshyn -- his fellow computationalist, but obviously the lesser pygmy).
“If thinking of teapots happened to be side by side in the brain with taking naps, maybe we would then revise our intuition that the two really haven’t much in common. But the issue is academic in the invidious sense since in fact there’s no good reason to think that similarity of psychological functions generally predicts similarity of brain locations or vice versa.”
ReplyDeleteI think this quote alone pinpoints accurately a problem that is seen quite often in the field of neuroscience, namely the confusion between the significance of results in the laboratory and in life more generally. Advances in neuroanatomy are impressive from the academic standpoint, and maybe, if it turns out that the hardware matters to the reverse-engineering of the mind, we’ll be able to use this knowledge to replicating a “thing that thinks”, but it would be a mistake to believe that the local activation of some part of the mind effectively means that this brain part is what brings about a brain state. Moreover, in my opinion, Fodor insists justly that we should not be too worried with those localisation questions when we take interest in the mind because we’re not even sure just yet whether the mind is localized at all. He does not talk about this, but I think all that functional brain imaging data could still be explained by a gestaltian theory, in which the local brain activation are just pieces that enable whatever the mind is to function. To link this with the article 4a, I think Fodor’s criticism would entice a whole different framework in talking about mirror neurons. The fact that a human integrates multimodal information, and the fact that smaller scale units like neurons also react to multiple types of information is interesting, but does not provide any more evidence that the localization of a neuron that fires when thinking about teapots and while napping a) gives meaning to this area of the brain and b) makes teapots and naps semantically, logically or behaviourally similar.
Yes, Fodor gave no alternative approach either.
DeleteHaving done EEG-fMRI research before, I share Fodor's outlook on how most of the neuroimaging research conducted on healthy individuals only show a correlation between a certain brain region and cognitive capacity without providing any new theoretical knowledge on cognitive functioning. I also agree that it proves to be a valuable tool in clinical research. But unlike Fodor I deem it to be useful in certain other areas of research. Neuroimaging research can be useful in understanding the developmental trajectory of brain development. Currently this capability is limited by the fact that it is extremely difficult to limit infants'/young children's movements during neuroimaging. However if we can figure out a way to limit their movements we can attempt to better understand the stages of brain development and corroborate it with neural data - further investigating ideas of sensitive periods, benefits of bilingualism, longitudinal impact of meditation training etc. The latter study can be conducted in adults as well - for studies looking at how certain practices or activities impact behaviour, their results can be corroborated with neural data.
ReplyDeleteAdditionally, if the research is well-designed it can be used to test innovative ideas. For instance, Dr. Rebecca Saxe has conducted research on individuals with normal vision and blind individuals to show that certain visual terms (such as flickering, twinkling) produce similar activation of visual brain regions in blind individuals as controls. This research questions our understanding of how these terms are learnt.
Apart from the above few aspects, I also find certain arguments in the paper ambiguous. For instance the sentence "if the brain does different tasks at different regions, that suggests it may do them in different ways" - although this is a possibility, it does not exclude the possibility of it being done in the same "way". I do not understand what "ways" stand for here - does it refer to similar algorithm, similar neurotransmitters or similar connectivity between regions. The phrase "different kinds of thinking" is also rather ambiguous.
How does "where/when" help explain cognitive development? (The Saxe work sounds like more correlations.)
DeleteI think I agree with Fodor on that neuro imaging can be abused in research. I totally see that finding out where the neurons are firing when certain mental states are going on can give people an illusion that we have acquired a very important knowledge about the mind – it effectively makes something intangible seem more tangible. I am not saying it is not important to know the location whatsoever, it has its own value of course (modularity), but in terms of understanding how and why a cognitive process works, it offers less than what it costs, especially if everyone does similar things and over and over again. For example, after an intense study of French, a (random, almost) part of the brain grows bigger than before, we know that this part must have been used, but then what does it tell us about learning French?
ReplyDeleteHowever, in everyday life, we do notice that, sometimes, the structure of an object gives us insights on how it works. (But maybe it is not a metaphor for the mind.) And I do find the Pavlovian story a bit too extreme – a lot of scientific discoveries are based on everyday observation, but not every observation is a scientific discovery.
Structure helps explain function when there is some similarity between structure and function, as in cardiac and other vegetative functions.
Delete“As far as I can see, it’s reasonable to hold that brain studies are methodologically privileged with respect to other ways of finding out about the mind only if you are likewise prepared to hold that facts about the brain are metaphysically privileged with respect to facts about the mind; and you can hold that only if you think the brain and the mind are essentially different kinds of thing. But I had supposed that dualistic metaphysics was now out of fashion, in the brain science community most of all. Brain scientists are supposed to be materialists, and materialists are supposed not to doubt that distinct mental states have ipso facto got different neural counterparts. That being so, why does it matter where in the brain their different counterparts are?”
ReplyDeleteHere Fodor asserts that neuroimaging studies are methodologically privileged only if you accept that through these studies you are merely discovering more about the brain- not the mind. In order to accept this reality, you must uphold some sort of dualistic thinking, which has been undermined time and time again by neuroscientist and philosopher alike.
Fodor argues that brain imaging cannot solve the easy problem of cognitive science, as finding out when and where the brain produces certain mental states does not address the questions of how and why we are able to think. However, I think that knowing what parts of the brain are active for specific mental states allows for a better understanding of how to reverse engineer some kind of thinking entity, whether that be a robot or a Frankenstein-like creature.
While I agree that some of the (publicly funded) research being conducted is not always necessarily very productive, I think there is something to be said about curiosity-based science. Even if the research doesn’t lead to an answer to the easy or the hard question of cognitive science, it can satiate some of the desire to understand more about our own minds (or as Fodor thinks, our own brains).
"Where/when" may not tell you "how" but it's not dualistic.
DeleteHow do you think it helps reverse-engineer cognition (or build a robot that can do what we can do)?
And how do correlations satisfy curiosity?
In this article, Fodor questions the importance being placed on researching neural localization and brain mapping in terms of understanding the mind. I'm mostly inclined to agree with his points that knowing "where" and "when" in the brain does not answer the hard problem of cognitive science, or provide a causal explanation of cognition. The problem is hard for a reason. I think neural localization and brain-mapping stems from our desire to try and make the abstract more concrete and visual. While I agree this doesn't seem productive in an overarching way of causally explaining cognition, I can see the desire for researching this.
ReplyDeleteBut, if the goal is to reverse engineer the mind, isn't the human brain a good place to look? Fodor does point out that "maybe we’re heavily invested in finding answers to which we don’t know the corresponding questions". However, I think that any observations or discoveries we make about the brain can lead to more and more insightful questions, so why fully dismiss this research?
In response to Anna, I agree with the idea that making more discoveries and observations about the brain could potentially give us further insight into how the brain works. By wanting to dismiss all of the localization research being done by cognitive scientists, like Fodor, what should cognitive scientists be researching then? We can’t try to reverse-engineer the brain without studying the brain. Although I do agree with Fodor that localization research is not going to help us solve the easy or the hard problem, I don’t believe that we should stop studying the brain. I believe that any new additional information we learn about the brain is good information to have. We are learning more about a system that we are trying to reverse engineer.
DeleteIf in the future, neuroimaging reaches a point where it can discern every single neuron in the brain and their firings, I'm not convinced that would get us any headway on the easy problem. Taking Fodor's point (when and where not telling us what and how) to the extreme, I still agree with Fodor. Maybe if we knew every detail of every neuron (when it fires, what neurotransmitters are involved, which neurons it synapses with), made a model, and tinkered with every combination of connections, we could solve the easy problem. That would be making a brain, messing with every variable to understand what affects what, and by the end, we should have an understanding of how and why we do what we do. The hard problem, unfortunately, remains untouched in this hypothetical. We'd always be blocked by the other minds' problem, though we could see the different behavior caused by each neuron being altered. I've been struggling with my next thought for the past two weeks, as it feels sacrilegious to the field, but I'm not sure the hard problem is solvable. Hopefully we'll read something later in the course that lights up a new path to the promised land, but right now, I feel like it's impossible to know what it feels like to be something else.
ReplyDeleteIf we figured out every detail of every neuron and made a model of the brain that works in the same way we do - would we not know how the brain works to do the things that we do? A comment was made in class about 'over reaching' to create a T4 before a T3 and how this could be a possible solution to solving how we can do what we can do, which has left me questioning the validity of this. It seems that if we can create a model of the brain (after we figure out the details and pathways of every neuron which correspond to a thought or action) and successfully create a robot (T4) that models the set up of the brain and it's neurons, then we will be able to grasp an entire picture that tells us how we can do what we can do. I agree that this would be grasping for something which is farther away (seems as though it takes more to build a T4 than to build a T3) but perhaps the only way to know how we can do things is to model it exactly after the system which allows us to do those things currently(?) I am still unclear where I stand with this but it seems possible to reach for T4 before or rather instead of T3. This may lead to understanding that it is this entire system (as a whole) that creates those actions, however, we may still not have an understanding of how individual processes are able to function and what exactly constitutes those actions.
DeleteWhile reading this article, I was somewhat relieved to learn Fodor’s opinions. As I was reading through article (4a), I kept asking myself how knowing when and where mirror neurons function in the brain and how they came to be is going to help cognitive scientists solve the easy and hard problem. After reading Fodor’s article, I was glad to see that my questions were not unreasonable. I agree with most of Fodor’s points although, I do not agree that we should stop doing or encouraging localization research (from a cognitive science stand point). Although I do not believe that this research will hep us solve the easy or hard problems, I do believe that it is valuable to increase the knowledge we have about the brain in general. If we stopped studying the brain, what else would we study to try and solve the easy and hard problem? I believe it would be strange to try and figure out how the brain works by ignoring the brain (trying to reverse engineer a system, without studying the system). Like Fodor, I cannot propose alternatives to study.
ReplyDeleteI enjoyed reading this because it reminded of how I used to think about things I didn't fully understand when I was younger because I didn't work with it personally. I understand why Fodor says what he says and I agree with him at times. For example, when he says, "what if, as it turns out, nobody ever does find a brain region that’s specific to thinking about teapots or to taking a nap? Would that seriously be a reason to doubt that there are such mental states?", I agree with this point of view, and many other points that he makes. However, there are quite a few things I disagree with.
ReplyDeleteFor example, at the end of the article, he suggests that he thinks this research may be a waste of time. From reading other skywritings, I learned that Fodor wasn't this biggest fan of neuroscience, and this view is evident from how he describes Pavlov's dog experiment. He dismisses Pavlov's experiment as if we only learned that dogs salivate when expecting food. We learned much more than that from this one study, and it influenced so many other findings that are useful in many other ways than just with neuroscience. It is important to note however, the study could have been done in many different ways, and without dogs or animals altogether. This finding was useful for answering different questions in neuroscience and other fields, but not so much for answering the easy question of cognitive science.
Although Fodor questions the usefulness of neuroimaging in reverse engineering (the easy problem), I think there is evidence where it is useful. For example, one study concluded that fMRI's give us the ability to answer questions concerning modularity of the brain, markers of different mental processes, exactly what information is held in what part of the brain, and theoretical questions in relation to the nature of tasks. (https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3610572/) Although there are questions neuroimaging could never answer for cognitive science, I don't think that it's a waste of time, and perhaps the usefulness of all this research will present itself in the future when there are more questions, answers, and information to work with.
Fodor argues that the how and why, not the what and when, should be the areas of focus when investigating mental functions. Along with this, he believes that there is no use in assuming that similarities in psychological function are correlated with similarities in brain locations. Resultantly, he challenges the academic overemphasis on neuroscience and dismisses its use of neuroimaging techniques / experimental data to address questions related to the mind. Related to my response on Cook et al, I agree with Fodor that purely attempting to localize mental functions to specific brain areas or patterns of brain activity does not give any leeway into the causal mechanisms underlying these processes. Additionally, this 'methodological privilege' of brain studies places a focus on neural behavior equivalence. As discussed in class, this overdeterminance does not simplify the easy problem and hard problem of consciousness (especially why / how do we know we 'feel'). Alternatively however, I conflict with him as I do believe that brain studies can give important insight into mental processes. For example, in individuals with brain lesions, investigating behavioural changes and localizing damaged areas with neuroimaging techniques have led to great leaps in understanding in cognition. With the seminal case of HM, we gained insights into the role of hippocampi in memory functions. While this does not give a complete picture to 'how' and 'why' memory is facilitated in these brain regions, I argue that this is still a more concrete start point for further investigation than the lack of alternatives that Fodor suggests.
ReplyDeleteIn Jerry Fodor’s piece, he challenges the methodology of the current practice of neural imaging the brain as a means of mapping neurological processes associated with certain actions. He explicitly challenges the importance of neural localization as a scientific endeavor.
ReplyDeleteIn doing so, he reveals that the scientific paradigm privileges an empiricism which demands that mental states be localized as a means of evidencing their presence, highlighting the futility of such actions, noting that we feel such states irrespective of their location in the brain. The how and why questions become far more important than the where and when which dominates current neurological emphasis.
I believe that neural localization does indeed have its uses – especially within the fields of neurosurgery and medicine. However, I do believe, that when it comes to answering the easy and hard problems of cognitive science, there is little use to localizing the specific neurons which fire in association to some specific action. Rather the why they fire, and the effects which they have, are questions that indeed need answering with regard to the mysteries of cognitive psychology
Before reading Fodor's article I did not question the usefulness or (un)usefulness of neural imaging and brain scanning. I did not think about how localizing parts of the brain which are responsible for the generation of certain acts, thoughts, ideas, etc., could be helpful in truly understanding how we come to perform those actions, think those thoughts, etc.. After reading Fodor's argument I am mostly swayed by his arguments. This is not to say that neural imaging and brain localization techniques are not useful. They most certainly are! Neural imaging is certainly helpful in clinical purposes: fields of medicine as well as studies about people who have lost either certain physical parts of their brain or lost certain abilities and how those two things are correlated. However, perhaps neural imaging doesn't bring us any closer to figuring out how and why it is that we do what we can do. Fodor's main point in this piece was that finding out 'when' and 'where' something occurs in the brain does not help us explain 'how' and 'why' the brain allows us to do what we can do. With this, I agree. If we figure out all the neural locations of certain actions, ideas, motor functions, thoughts, etc., we are still not much closer to finding out how we have those actions and thoughts. If we figure out the part of the brain which is activated when we think of 'dogs' for example it will not tell us how exactly we are able to 'think of a dog'. Additionally, if we are to try to create a robot (T3) that can do everything we do, the materials used to create this robot will likely not be the same as what our brains are made out of (this would be more like T5) anyways. Therefore, knowing the physical part of the brain will not help us in being able to build the robot to do what that part of the brain is doing - which is still very unclear even if we know which exact part is doing such a thing.
ReplyDeleteBut why (unless you’re thinking of having it taken out) does it matter where in the engine the carburettor is? What part of how your engine works have you failed to understand if you don’t know that?
ReplyDeleteI usually don't have a quote that jumps out at me but this quote in particular I think sums up the last lecture pretty well. Over the last few weeks we have gone on a journey first learning what computation is and how it aims to solve the easy and hard problem. Then we learned who Turing is and whether he was considered a computationalist. He doesn't seem to agree that computationalism can necessarily solve the hard problem but he does seem to be confident that it can solve the easy one/ get at behavioural equivalence. We then moved on to judge whether the turing test could be solved by T2 and came to the conclusion that T3 may in fact be needed to do the turing test forever with another human. With that said, we moved on to why T3 needs to be T3 and why T4 may not be necessary. I think the quote above seems to sum up Fodor's opinion at least. He definitely does not think T4 is necessary and worth pursuing. T4 requires the right thing to be in the right place and for it to perform the same thing that it does in the brain biologically. However, as Fodor says, what does it matter where a specific part of the mechanism is? It doesn't necessarily matter to the general nature of how a mechanism works. Now he does go on to say that there may be some clinical pros to knowing where things are for the safety of the patients going under the knife, however, when it comes to our understanding of cognition, I think Fodor stands on team T3 and not T4. As long as we know where the function is conducted, the organization stands to be irrelevant. As fodor clearly implied in his article, it doesn't matter where the components are as long as you know they are there. As long as you know there are mental states happening, why does it matter where they are happening in the brain? Fodor gives the last string of hope for cognitive neuroscience to serendipity. However, he immediately bring that argument down as well as serendipity is really not a strong basis to conduct science on.
In-class we’ve repeatedly been challenged about how experiments seeking to explain cognition and other phenomena of interest often run and collect data before they have a clear idea of what it is they’re collecting data about. It’s been suggested that, although their findings may loosely relate to cognition, they do not specifically answer the elementary questions of how and why processes function the way they appear to. Fodor also addresses this as he criticizes the excitement surrounded research aimed at locating neural firing associated with various activities (“functional localisation by neural imaging”), with the assumption that the brain is organized according to geographical principles. However, I wonder if we’re too quick to dismiss the value of these improvised experiments. When we have an idea, even if we don’t have a perfect definition, empirical research psychologists’ first impulse is to try to measure it. But perhaps there may be a method to the madness of trying to create empirical experiments pertaining to yet-to-be-defined concepts, because it may be one way to refine our understanding by acquiring more information and thus ad hoc definitions. It seems to me that much of neuroscience does operate according to the assumption that acquiring new data inherently fuels progress by contributing to a fuller picture of reality. While I take Fodor’s point that we may be sipping a step in not explicitly stating the reasoning behind these assumptions, I don’t see Fodor offering an alternative approach. Blind trial and error based on the limited information available seems to me like the most basic description of what it is to do scientific research.
ReplyDeleteIn the dairy, Fodor is skeptical of the overwhelmingly popular trend in Times, or in academia, of neural imaging studies of specific mental activities and of the implications of this pursuit. In the passage, he said: “As far as I can see, it’s
ReplyDeletereasonable to hold that brain studies are methodologically privileged with respect to other ways of finding out about the mind only if you are likewise prepared to hold that facts about the brain are metaphysically privileged with respect to facts about the mind; and you can hold that only if you think the brain and the mind are essentially different kinds of thing.”
Here, Fodor claims that the only if people think that brain and the mind are distinct beings will someone think that the facts about the brain are superior metaphysically to that of the mind.
I wonder why this is the case for Fodor. From my best knowledge, it is not inconsistent to have a non-dualist who think mind and the brain is inseparable and the facts about them have no priority over one another, yet because this brain-imaging operation is something that we can observer and infer on directly according to the technology we have for now, so the brain study is privileged. The subject of the study here is not separated from the mind, but it is the observable and quantifiable part of the brain. This rationale could also make sense, as Turing would probably agree, you investigate on what you can observe.
“But it’s functional localisation by neural imaging for which the Times is especially enthusiastic; and I’d guess that as the Times goes, so go the grants. It particularly likes those polychrome maps that show a place in the brain that’s red when you’re thinking about one thing and green when you’re thinking about something else. (Disappointingly, I gather it’s not that the brain turns red or green depending on what you’re thinking about; the colours are computer generated to summarise the levels of neural activity that the experiments discover.) Well, to come to the point, I wonder why the Times cares. I wonder why anybody cares.”
ReplyDeleteWhile a lot of scientific articles about the brain brought into popular culture serve as “click-bait” or present something flashy that might not be fully supported, I do think neural imaging is worth investigating. I understand that even if we could map out all the things we think about to “where” in our brains we think about them, the question of “okay..so what?” is still not answered. I see localising brain functions less about mapping and more about understanding cognitive functioning through connectivity and dynamicity. For example, if you saw that while someone recalled a memory of how to solve a calculus problem, areas in the brain that were attributed to olfactory senses also “lit up”, then maybe it would change how you went about learning calculus? Seeing a potential system of cognitive function through active brain areas could maybe help unpack the faculties required for higher order cognitive functions.
This comment has been removed by the author.
ReplyDeleteI agree with you on that “if different tasks map to a same location, the brain has probably found a unifying strategy (i.e. one function) to solve them. Similarly, if one task maps to different locations depending on the context, the brain likely factors the task in smaller functions than we first thought. Thus, over time, we can build a model of how the brain partitions function-space just by looking at where particular functions are located in the brain”. But based on this, it seems that we do not need to know where exactly the functions are, we only need to know that such functions exist. Also, I do not think the function-space mapping necessarily simplifies the reverse-engineering problem. The brain does not always do things in the most optimal way. Reverse-engineering only needs to provide a doable solution, not the exact solution that the brain uses. I do agree that, to some extent, “function-location studies do seem to inform us on how the brain functions”. But I think simply with function-location studies, we cannot know “how the brain functions the way it functions” causally.
DeleteFor the past three years at McGill University taking classes in Psychology, Cognitive Science, and Neuroscience, I have yet to question the validity or importance of neural imaging of cognitive structures. Fodor eloquently explains the significance of questioning the ‘why we do what we do.’ The when and where mental states occur leads us to a dead end. It wastes tremendous amount of time and money. Counterexamples have been given throughout the sky writings of vegetative structures, but that strays from Fodor’s argument. In a clinical setting, neuroimaging has helped to reverse engineer structures like the cochlear. Fodor agrees. However, the focus of Fodor’s paper is how does neural imaging help to explain the how and why of cognitive functions. I have been scouring the internet for an example of when neural imaging of mental states tells us (leads to) how and why these mental states occur. Sadly, no luck. Has anyone else found an example? I do wish Fodor offers some guidance of where the scientific community should head.
ReplyDeleteRe: Why, why, does everyone go on so about thebrain?
ReplyDeleteIn this piece, Fodor explores the question of how and why the brain produces a certain cognitive capacity. Fodor mentions that to understand an engine, it is sufficient to know it's function as a whole rather than just the localization of the parts (which is an explanation analogous to how we should understand brain). I find myself agreeing with this statement.
To put this into perspective, I personally don't think brain imaging research should be limited to surgery interests of just “taking out parts”; there must be an understanding of the representation in the brain to do so. It is important to look at pathways and relationships between cortical areas. If you performed brain surgery without an understanding of how the brain works as a whole, you might end up damaging sensation to your hand or speech areas, or even change your personality.
However, I do not agree with his claim that neuroimaging won’t get us anywhere in solving questions regarding cognition. If we stopped studying the brain, what else would we study to try and solve the easy and hard problem? I cannot imagine trying to reverse engineer a system, without studying the system… although, I cannot think of other alternatives to study to get to relevant answers.
“ I admit, for the sake of the argument, that consciousness is correlated with certain neurons firing at 40 Hz cycles; and that some bits of the brain light up when we hear nouns but not when we hear verbs; and that there are (different) bits that light up when we see a thing, or form its mental image, but not when we hear a thing or describe it to ourselves. It appears there’s even a place in the brain that turns on just when we hear a word that stands for a vegetable; ‘lettuce’ excites it but ‘roast beef’ doesn’t. So be it.”
ReplyDeleteI agree with Fodor here. Knowing where and when doesn’t tell us how and why. We know the fundamentals of how the brain works, it fires action potentials, neurotransmitters are released in synapses, chemical reactions are generated etc. this is the biological, physical level of how the brain works which doesn’t tell us anything about how the mental states arise. In reverse engineering, we are trying to reproduce the same outcome; it doesn’t matter how it is done. The only way I could see it be useful is if in the future we are able to completely reconstruct a brain, then it would be interesting to know how the synapses are organized in neural networks to produce mental states like memory for example. We could then reproduce these neural networks in order to generate the corresponding cognitive performances. I think that to reverse engineer cognition , observing the outcome of what we want ( memorizing for example) and understanding which processes and mechanism humans use to do it ( we form categories, associations, use mental imagery) is what we enable us and is enabling us to replicate these functions in AI for example.
“To put the same point the other way around: what if, as it turns out, nobody ever does find a brain region that’s specific to thinking about teapots or to taking a nap? Would that seriously be a reason to doubt that there are such mental states? Or that they are mental states of different kinds? Or that the brain must be somehow essentially involved in both?”
ReplyDeleteI like how this part puts things into perspective because it’s true that we often what to know what each part does within the tiniest details, but the way it does what it does not mean there are no mental states. The geographical situations of the parts in our brains do not really give us an explanation of how it actually works.
The author starts by questioning why there is such a fascination with the brain. For me personally the curiosity originated in understanding people’s actions, reasoning, motivations and ultimately the changes and events (in physiology and environment) that can negatively impact these abilities.
ReplyDeleteA point of contention in this article for me is the mention of competition for resources. While I understand that coming from someone who personally is not interested in neuroscientific research this may be difficult to see, there are so many benefits and other areas of impact to neuroscientific research. For one, research on the brain could lead to medical discoveries which (continuing from the authors point of view) in the long run would decrease the expenses of medical care along with other industries. Depression, for eample, is the leading cause of disability worldwide and the estimated economic burden is estimated to be $210.5 billion USD a year with half of these costs being attributed to work absence and decrease productivity, the other half being medical expenses.
He mentions whether or not mental functions are neutrally localised in the brain. While it connects with the philosophical question of how the mind works as he mentions, a simpler way of viewing this I believe is from the angle of pathology. While certain mental health issues have shown correlations with different brain regions, alternate hypothesis of hormone disfunction have been presented. The link between the brain and character distortions due to illness or injury are harder to contend with: Traumatic brain injury patients show differences in character due to brain damage. This cannot be questioned.
While it is true that the reason for the rise in neuroscientific research is due to technological advancement, it is not the reason that the research has been so strongly driven, it has simply made the research possible. The benefits of knowing more about the physiological functioning of the brain, whether or not the questions on the way that thought works or whether there is a separation of mind and brain are fully understood, are insurmountable.