Dear Joaquin,

I disagree as this concept suggests that decisions are made in the frontal lobe, and the rest of the brain and body slavishly executes whatever command is sent from “central headquarters”. So in my view, instead of keeping dualism at bay, it introduces a new kind of dualism (or even a homunculus to the malicious reader). More embodied approaches to cognition emphasize the dynamic interaction between brain parts and the body in the creation of behavior. I have written a paper on the origins of behavior with Aaron Schurger (attached). Also, I’ve written a paper with Dan Burnston and Pim Haselager (in which we build on your work on monkey physiology) that frames the role of the frontal lobe in terms of regulating autonomous sensorimotor and decision processes, rather than the initiator of behavior (also attached). Maybe your suggestion is compatible with these considerations, but at face value it suggests a more amodal, detached and disembodied framework.

Best regards,

Sebo

PS. Apologies fore the shameless promotion of my own work, but I believe these papers are relevant in this discussion.

Article Why we may not find intentions in the brain

Article Nowhere and Everywhere: The causal origin of voluntary action

Dear Joaquin,

I disagree as this concept suggests that decisions are made in the frontal lobe, and the rest of the brain and body slavishly executes whatever command is sent from “central headquarters”. So in my view, instead of keeping dualism at bay, it introduces a new kind of dualism (or even a homunculus to the malicious reader). More embodied approaches to cognition emphasize the dynamic interaction between brain parts and the body in the creation of behavior. I have written a paper on the origins of behavior with Aaron Schurger (attached). Also, I’ve written a paper with Dan Burnston and Pim Haselager (in which we build on your work on monkey physiology) that frames the role of the frontal lobe in terms of regulating autonomous sensorimotor and decision processes, rather than the initiator of behavior (also attached). Maybe your suggestion is compatible with these considerations, but at face value it suggests a more amodal, detached and disembodied framework.

Best regards,

Sebo

PS. Apologies fore the shameless promotion of my own work, but I believe these papers are relevant in this discussion.

Article Why we may not find intentions in the brain

Article Nowhere and Everywhere: The causal origin of voluntary action

Calling the frontal lobe "the executive of the brain" might be a bit too restrictive even though damage to this anatomical part of the brain is highly correlated with diminishing executive function. But, beyond anatomical location, we want to know the neuronal mechanisms that enable executive planning and decision making. For a description and discussion of some candidate brain mechanisms see "Composing Behavior: Registers for Plans and Actions" on my RG page.

With due respect to all findings that operation or failure to operate of the frontal lobe do in fact change executive function, I cling to the view that executive functions are not located in this or that part--or even located strictly in the brain. Rather, I believe that they follow from the material hierarchical organizations that spread across many scales of space and time all at once. After all, it is those many scales of space and time that we often presum executive functions to operate upon.

For instance: http://www.ncbi.nlm.nih.gov/pubmed/24628057

This view is not at all antagonistic to the wide swaths of evidence in favor of implicating frontal lobe in executive functions, but what's unfortunate is that many of the authors presenting that widespread evidence are just plain ignorant of what else might be out there to support their theory-building.

Best wishes,

Damian

The terminology is good enough. Regarding the decision being taken before it enters consciousness, I believe that all these experiments are flawed in the sense that, after the "consciousness" part is materialistically described in terms of a signal travelling through a set of processing areas to the speech-generating center, the results can be explained in terms of processing delays and uncertainties; "consciousness" is not needed for these analyses. I do not think that any experiment can address the subjective part of consciousness; we need to stay within the materialistic framework. 

Joaquin, I believe your own approach to this issue has gone some way to downplay the role of a central executive in the prefrontal cortex by accentuating the distributed nature of cognition, especially with regard to memory.  An approach to which I am sympathetic.

An extract from one of your papers may be relevant here:

Fuster, J.M. 2009. Cortex and Memory: Emergence of a New Paradigm

Journal of Cognitive Neuroscience 21:11, pp. 2047–2072

"Converging evidence from humans and nonhuman primates is obliging us to abandon conventional models in favor of a radically different, distributed-network paradigm of cortical memory. Central to the new paradigm is the concept of memory network or cognit—that is, a memory or an item of knowledge defined by a pattern of connections between neuron populations associated by experience. Cognits are

hierarchically organized in terms of semantic abstraction and complexity."

While this article does not address the concept of "Executive of the Brain" specifically, it does show how to think clearly about ANY concept related to the brain.

https://faculty.franklin.uga.edu/rkthomas/sites/faculty.franklin.uga.edu.rkthomas/files/ManuscriptVersionPostEdit.pdf

I do not know if Dr. Fuster is questioning the "evolution" of his own thinking, or sampling the views of others - perhaps both. I have seen his hypotheses move what I consider "forward," especially in his 2009 and 2015 publications. I do not think the thoughts in these publications are necessarily exclusionary to any of these RG answers. However, these hypotheses do not go far enough. There seems to be an assumption that neocortex and frontal lobes "run the whole show." But the executive function of working memory is quite dependent upon the "subcortical" Basal Ganglia, while from a phylogenetic perspective, the bottom-up development of the frontal systems, or the frontal-striatal-pallidal-thalamo-cortical circuitry system seems a much more biologically consistent way to understand executive functioning; there is a vast literature that describes this system. In fact, perhaps I incorrectly assume all on RG know of the article by Alexander, DeLong, and Strick (1986) which first spoke of this circuitry, and has become the most quoted paper in neuroscience! This "system" controls access to working memory, described in a paper circa 2008, by McNab & Klingberg. Similarly, the phylogenitically "old" cerebro-cerebellar circuitry system is also fundamental to the support of EF, described by Schmahmann & Pandya, circa 1996. The conclusions that the Basal Ganglia and Cerebellum are merely co-processors of movement are ancient and, IMHO, should be buried. Anyone who searches through RG will find numerous papers and reference books that support the conclusion that "EF" is hierarchically organized in the vertebrate brain and is an ongoing expansion of the cortico-striatal and cerebro-cerebellar systems, which operate as an integrated unit. It does not seem tenable to consider the prefrontal cortex as "king of the mountain." "EF," and its so-called subdivisions, all need to be reconceptualized, even from this relatively "basic" understanding of the integrated brain.

The picture has been further changing for two reasons. First,network or "systems science" has been applied to the study of brain-behavior relationships. Second, technological advances, including Bayesian statistics, have been applied to brain-systems/network science. This has generated an entirely new model of brain-behavior relationships that is highly dynamic and flexible, simultaneously supporting the view that neocortex, basal ganglia, and cerebellum define the integrated brain, calling for new practical test paradigms. For instance, many of the assumptions inherent in the artificial construct of "an executive brain" are nicely challenged by Sebo's and Damian's answers, as well as others such as Cisek, Kalaska, Pezzulo to name just a few. Many of Dr. Fuster's earlier investigations were based upon a "perceive-think-respond" experimental paradigm; however, the vertebrate brain evolved to serve the needs of interacting with an ongoing, dynamically changing environment; this requires continual neural processing, very consistent with brain-systems thinking, concepts of embodied thinking, decision-making that is based upon activation over large regions of cortex and decisions made in "real time", and a viewpoint of "EF" that accounts for "bottom-up" support without any ongoing need for top-down control and decision-making. Finally, especially to Sebo, there is nothing self-promotional about posting one's own relevant work, there is no need for apology about such posting, and the fact that I have chosen not to attach my own work says nothing about the quality or usefulness of this reply.

Len Koziol

Dear friends,

            Many thanks for your thoughtful and stimulating answers to my question.  Thanks also for your appropriate papers and references.  All of your comments, without exception, are generally in accord with the views I have expressed elsewhere, some of you contributing interesting qualifications to them.  As most of you guessed from my rhetorical tone, especially if you knew a little about my work, my answer to the question is that the concept of an executive in the brain is not at all useful.  In fact, for all its metaphorical appeal, it is a regressive concept. Here is why.

            1.  Conjuring in the brain a central executive agency inevitably leads to an impasse similar to the one that philosophers Descartes and Kant encountered with their supposed “spontaneity” or “autonomy” of the mind. My rhetorical statement that the concept keeps dualism at bay is utterly false.  Yes Sebo, it makes dualism worse, by multiplying it!  In fact, it leads to an infinite regress: who or what controls that executive?; and then that other one?: and that other one? And so on…ad infinitum, a never-ending hierarchy of executives, in the brain or out of it.

           2. The concept of a central executive in frontal cortex takes this cortex out of the perception-action (PA) cycle and of the enormous mesh of interactive neural networks that dynamically adapt the organism to its environment.  Action can begin anywhere in that cycle.  Yes Len, the lateral thalamus-basal ganglia loop described by Alexander (my graduate student) et al. is embedded in the PA cycle and very much involved in its functions.  You say,  “…[T]he vertebrate brain evolved to serve the needs of interacting with an ongoing, dynamically changing environment; this requires continual neural processing, very consistent with brain-systems thinking.” This, in my book (literally), is the PA cycle.  Reading your comment it seems like I am reading my own stuff!   Let me just tell you that I cannot say it any better than you.  As for Bayesian probability, I think it permeates all cognition.  To wit: Fuster and Bressler – Past makes future: Role of pFC in prediction. J. Cogni. Neuroscience, 27: 639-654, 2015; and forthcoming article with Haykin, on risk control, to appear in IEEE Access.

     3. Addendum: I don’t want point #2 to be misunderstood, as it seems to me    possible when I re-read it. When I say that the concept of an executive in frontal cortex takes the latter out of the PA cycle, I mean that that concept theoretically gives the frontal cortex a supra-ordinate function that overrides that of the PA cycle, which is demonstrably not true. On the contrary, the frontal cortex is very much a part of the cycle, and, within it, the prefrontal cortex in particular, coordinates the executive functions of executive attention, working memory, decision-making and inhibitory control; all of which it does together with other cortical and subcortical structures. In that sense it is appropriate to call the prefrontal cortex “executive cortex,” but without extricating it from the PA cycle or treating it as a sort of executive “homunculus” dishing out the orders to every other part of the brain. Four decades of working on the prefrontal cortex have taught me at least that much

 Thanks again to all.  But, humbly, I cannot pretend to have the last word on this!

Joaquín

Dear all

 A recent article "Banishing the Control Homunculi in Studies of Action Control and Behavior Change" written by Verbruggen et. al.(2014) proposes a new perspective on executive control system. Even though we accept that there is executive centre in some part of brain does not explain how control is achieved. Control occurs at different time scales in different tasks so a unitary control system does not fit. There seems a need to deconstruct the control system in its constutuent  components (perceptual, cognitive ,motor).There have been slogan of “Dissolve,deconstruct, or fractionate, the executive! Let a hundred idiots flourish!” by Monsell and Driver (2000).

in line with the decomposition of executive, Salinas and Stanford,2013 suggests that detection of stop signal is critical for successful stopping in countermanding task. The presentation of stop signal decelerates the Go process and that results in stopping. Role of detection, attention, rule learning etc can not be ignored in control. Recent study on ADHD shows that the response inhibition impairment may be due to perturbed attentional system.

Dear Indrajeet,

“Even though we accept that there is executive centre in some part of brain does not explain how control is achieved. Control occurs at different time scales in different tasks so a unitary control system does not fit.”

         You are entirely correct.  An executive center without control mechanisms is an empty neural entity.  As far as I am concerned, you can place the executive anywhere in the brain, or even call the entire brain the executive, that is, the captain of the ship.  You still need the steersman, the agent that navigates the boat with the appropriate action at the appropriate time, as the signals from sextant and wind direction require. The key is time, as you imply, and the parsing of the control mechanisms in the time domain.

         In my view that is precisely the role of the prefrontal cortex, sitting on top of the PA Cycle and guiding the actions to their goal with cognitive control of attention, working memory, inhibition, and decision-making.  It is by control of that cybernetic cycle (from Greek, κυβερνάω, to navigate), with feed-forward and feedback mechanisms, that the prefrontal cortex mediates cross-temporal contingencies and controls behavior and language.  But you do not need to place the “executive” there, and the cycle runs not only through the environment and the prefrontal cortex, but also through other cortical and subcortical structures.

J.M. Fuster - The prefrontal cortex-An update: Time is of the essence.  Neuron, 2:319-333, 2001

         With regard to your comment on the “stop signal” and its role in attention, I am persuaded by the evidence, summarized elsewhere, that ADHD is caused by a laggard maturation of prefrontal inhibitory control in childhood and adolescence.

Cheers, Joaquín

in a clinical setting "executive brain function" and "prefrontal cortex" activity can be a very useful concept when translated in terms such as "Leadership" and "Self-Management," and "taking charge of your life, your time, your old lower brain habits."

I've even used it with children diagnosed with ADD and ADHD. Watkins and Watkins, Maggie Phillips, Claire Frederick and others have written of "Ego-State Therapy" and putting the Executive in charge of integrating personas and lower-brain functions around a leadership mission, values, and objectives.

Splendid, Neil.  I also have an interest, and done some work, in the training of executive prefrontal functions, especially working memory, in education and TBI. Now with Wright we are trying to prime them with anodal tDC.  Cheers, Joaquín

Hi all.

If any of you are interested I am a visual artist and within the phenomenon of vision it's possible to identify three protagonists Vision-Space: The Protagonists http://youtu.be/516mjrU3aC0. (a) Implicit awareness that is prior-to or in temporal advance of, (b) conscious  awareness (explicit). (c) Mind that has to mediate between them based on intent in the world. Awareness would appear to be something of a dance that would involve the entire brain? Vision-Space: Awareness is a dance mediated and augmented by ‘mind’ within a multidimensional space https://youtu.be/-DCx5kLS2MQ

Attached is a list of my musings, as that is all that they are. Initial Vision-Space technology has been created on the basis of these musings however!

Best - John

Joaquin,

Generally speaking in science, concepts are useful for explanation and/or  prediction. As a practitioner of translational medicine I am mainly interested in the predictive power of a notion of EF. Accordingly I have introduced an intensional definition of  EF (Labra-Spröhnle, 2015,2016); paraphrasing Anokhin (1974): EF are any of “those specific mechanisms of the functional system which provide for the universal physiological architecture of the behavioral act.”

Please see attached papers with my detailed answer to your question. I hope this can be "useful".

Cheers,

Fabian

Article The Mind of a Visionary: The Morphology of Cognitive Anticip...

Article Human, All Too Human: Euclidean and Multifractal Analysis in...

Article Restoring the integrative value to the notion of executive f...

Dear Fabián:

             Thank you for very interesting papers.  To begin with, a little defensive statement is here in order.  As you may have seen in our discussion, my dismissal of the neurobiological concept of a central executive does not at all extend to executive functions, especially of the prefrontal cortex, to which I have dedicated much of my scientific life.  I am even willing to concede some heuristic value to the concept of a central executive. 

             I love your first article, in Cognitive Systems Monographs.  I am a great admirer of Luria, whom you extol.  I did not have the pleasure of meeting him (though I met Anokhin).  Luria was certainly a visionary, and a great clinician.  The truth is that he shaped profoundly my thinking with regard to the role of the prefrontal cortex in language, the temporal organization of behavior and, above all, the idea that you so well describe: cognitive anticipation.

             Indeed, if all the executive functions share one characteristic in common, it is their temporally prospective aspect; including working memory (see the discussion I initiated with my question: “Why is working memory considered a special kind of memory instead of a state of memory?”)

            On a related subject, I am sure you recognize the influence that Anokhin and Bernstein have had in the development of my concept of the perception-action cycle.

 Saludos amigos!

 Joaquín   

I personally think that an "executive mechanism" is a very useful concept. However, as pointed out by Sebol, this concept introduces dualisms: process vs. data, top-down vs. bottom-up / feedback vs. feedforward, source vs. site (of attention effects) etc ... These are all very useful dualisms. They have been used in numerous studies, and are recurrent topics in the psychological sciences.

To me, the question is not whether an executive system is a useful concept, but rather, where are the boundaries of this metaphor? I am a big fan of Joaquin's elegant solution to define the sulcus centralis (Rolandic fissure) as the anatomical boundary between the executive system (in prefrontal cortex) and sensory systems (in posterior cortex).

I however wonder whether it makes sense to attribute executive functions to some poterior regions? For example, the intraparietal sulcus (IPS) in the posterior cortex has been recently linked with control functions of WM (e.g., Cowan 2011, Neuropsychologia)

Furthermore, attention effects are often observed in sensory areas, such as visual cortex during WM or perceptual attention tasks. Can we maintain the simple distinction between the source of attention effects (in PFC) and the site where these effects manifest (sensory cortex)? I mean, how is attention "implemented" in sensory cortex? Do we have to assume that sensory cortex has its own control mechanisms, to bias information at the site where it is represented?

cheers

Tobi

I concur that the notion of executive functions is a knotty one. I have conducted over 500 frontal lobe examinations utilizing and 11 test battery on individuals from 5-75 and find that 60% of them cannot pass with a perfect score  regardless of age or education. I note with respect to clinical evaluation and additional neuropsychological assessment that this 60% of my population have decided difficulties both overt and subtle, with decision making. Whether the frontal areas are the "site" of executive functioning or simply a hub in the network may be irrelevant from a clinical perspective. 

Dear Tobias,

Yes, you are correct, the concept of “executive mechanism(s)” under prefrontal control is a valuable, I would say essential, concept.  Those comprise the mechanisms of attention, working memory and decision-making.  Those executive mechanisms, however, are part of the perception-action cycle, which circulates through the cortex (both frontal and posterior) and the environment, with feed-forward and feedback.  Of course, sensory—posterior--cortex is also part of the cycle and the mechanisms, and therefore it participates in attention, working memory and decision-making, together with the prefrontal cortex.  Remember, the prefrontal cortex does not do anything by itself.  It needs all the cortex and assorted subcortical regions to do the job.  There are the constraints to the metaphor of the “executive” that you are asking for.  Cheers, Joaquín   

Dear all,

I just had another glance at all posts on this thread, and it seems that we generally agree that the executive system is a useful concept, although there were also some comments on its limitations. There is further consensus that the PFC is on top of the hierarchy of executive control.

Let me ask a related, but trickier, question:

"Resources of WM": is this a useful concept?

Background: It is well-known that WM has a limited capacity, and there is a lively debate about the nature of the "resources" underlying this capacity limit. Frankly, I am not sure whether resources is a good scientific concept. For example, it is not clear whether we are talking about brain structures (e.g., the PFC), cognitive processes (e.g., attention) or about the energetic "quanta" that fuel these cognitive processes (metabolism).

It is further not clear whether resources refer to executive control or sensory mechanisms, both of them combined, or even something that is more than the sum of these parts (compare connectionist accounts, e.g., Fuster, 2009, JoCN).

As mentioned by Joaquin, cognitive functions such as WM or attention are the result of an interaction between executive and perceptual neural networks in the context of the perception-action cycle. Hence, stating that attention - or if you like, executive control - operates in a strict top-down fashion is only half the truth ... citing Joaquin: "The PFC does not do anything by itself."

A question I find particularly interesting is whether WM in different modalities relies on a shared / central resource (as claimed by Cowan, 2011, Neuropsychologia), rather than distinct resources (compare Baddeley's model). Since WM emerges due to recurrent reverberations of feedforward and feedback signals between posterior and prefrontal cortex, it does not make sense to attribute WM capacity limits to one type of resource without taking the other into account.

So, what do you think about the resource concept?

Is it useful, or does it raise more questions than answers?

cheers

Tobi

PS: Pardon me for this attempt to hijack of a thread I didn't even start. 

I am just curious about your opinions!

Aloha,  

When Pribram mused about the executive of the brain in the late 60's, he had NO useful data about the function of the Cerebellum - it was too dense to penetrate with any resolution at that time.  Indeed, the Cerebellum has only recently yielded SOME of its dense mystery. My experience suggests that the frontal lobe PLUS the cerebellum might be a useful candidate for 'executive of the brain'. Note how often that Coup-ContraCoup injuries are implicated in studies of the effects of frontal brain injury; such injuries often involve the Cerebellum. Prejudice re Cerebellar function, plus aforesaid lack of scan technology, has lead to sparse assessment of this essential conductor of brain activity.  

I've been entranced by the complex question of assessing Executive Functioning (EF) for 40 years in clinical settings. Experience in Psychiatric and Correctional populations suggest that there are two domains of EF: concrete-literal thinkers, & gifted, overexcitable thinkers. Bright concrete thinkers, who are unconflicted by emotional issues, do well on go-no go tasks, less well on abstract tasks. Gifted thinkers show lower scores on concrete tasks that factor highly on reaction time, and higher scores that are abstract & challenging. See the work of Linda Silverman (eg, Asynchrony), and the wealth of papers examining the non-normative scores of gifted people, such as: Gallagher, S. A. ( 2013). Building bridges: Research on gifted children's personality from three psychological theories. In C. S., Neville, M. M., Piechowski, S. S., Tolan, (Eds.) Off the charts  : Asynchrony and the gifted child.(pp.48-98). Unionville, NY: Royal Fireworks Press.

In my experience, this body of work persuades a reasonable researcher that EF needs to be researched in the context of sub-groups: it does not fit a bell-curve assumption. Once this caveat is honoured, the confounded data in papers researching EF will make much more sense.              Good fortune in your work,   Psychologist Paul McGaffey PhD(ABD)

@Tobi,

You are not hijacking! To start: is a concept really nonuseful if it raises more questions than answers? The concept of attentional resources is probably connected to that of working memory, and I would be skeptical of any model that started at one and didn't wind up at the other. Your text references indicate their association since the '80s. Even disregarding those points, there are good philosophical arguments for the prima facie utility of concepts that generate more questions than answers, i.e. solemnly intoning "God save the queen" with little evidence for the existence of either!

Joking aside, I am not sure I understand whether you are saying (1) 'PFC is not the only game in town' or (2) that there is some kind of Reynold's number coming into play when you are talking about "metabolic quanta" that underlie WC processes. I think my rephrasing of your interrogation is shabby, and I apologize for that, but to me your discussion of things that aren't clear is itself unclear. Specifically, I would like to know: What resources would you hypothesize for the capacity limits of working memory?

This is a very interesting question to me.

There is one more comment that it is necessary to make. Someone once said (Lanier?) that, as far as computers are concerned, "the file is likely to last longer than the photon." Their reasoning was that the fundamental particle of light was more vulnerable to being theoretically superseded -- paradoxical at first glance, but the point was about the subtlety of physics and models (compare the happy situation of today's neuroscience, and contrast the practical necessities of instrumenting a read-write system in a CPU). Does this metaphor shed any light on your difficulties?

@Len Koziol,

I am very impressed that you detected the rhetorical device in the opening post by Dr. Fuster. But do you agree with his assertion that Bayesian approaches can cover or even permeate the entire hierarchy of cognition? I intuit a few unprojected predicates here.

Thank you all,

CLE

@Charles: Hi there, and thanks that you commented on this question - I already had begun to believe that I am the only one concerned about the usefulness of the resource concept.

First of all: Oh yes I certainly agree that the PFC is not the only player in the game of WM. Secondly, apologies if my description of the resource concept was somewhat vague. This is partly because I do not know what resources are really. Different authors have different ideas about the resources underlying WM capacity, although there is some consensus that attention plays a role.

Two important points about the resource concept:

a) Resources are limited. They are placeholders for some mechanism that is capacity limited (a cognitive bottleneck).

b) Resources can be continuous or discrete

The quantity and quality of WM representations have been linked with discrete and continuous WM resources, respectively. For example, the observation that humans can maintain about 3-4 items in WM has been attributed to a discrete resource, such as a fixed number of 'slots' (Zhang & Luck, 2008 Nature). These slots are like drawers in a cupboard. A fixed resolution slot model posits that each slot can hold one item, regardless of its complexity. Even though copies of the same item can be put in different drawers, it is not possible to store two items in one drawer. In contrast, continuous resources do not come in such quanta like drawers, but can be flexibly distributed between items. The more resources an item receives, the higher the precision of the WM representation (for details, see Ma et al., 2014, Nat Neruosci).

Curious that you mention light photons, because of the dual nature of light: it is both particle and wave. Similarly, WM capacity appears to be limited by both types of resources: continuous & discrete (e.g., Xu & Chun, 2006 Nature). So far so good.

The point that bothers me is that the resource concepts allows for describing how WM ‘behaves’ - without explaining the underlying neural mechanisms. Attempts have been made to link discrete WM resources with certain brain areas (such as the intraparietal sulcus, IPS: Cowan, 2011 Neuropsychologia), but the idea that our brains have a limited number of slots / drawers is not very appealing from a neuroscientific perspective. More plausible is the continuous resource concept, as it has been linked with neural noise (p 347ff in Ma et al., 2014) or competition for 'cortical real estate' (that is, for representation in two-dimensional maps, which can be found in several brain regions: Franconeri et al., 2013 TiCS).

In particular, I am not sure whether the resource concept is of any use from a connectionist perspective: WM emerges due to the conjoint activation of neural networks in distributed brain areas that interact continuously (for details, Fuster, 2009 JoCN). If this interaction is critical for the attentional activation of memory representations - what can we state, except from the trivial truth that WM relies on distributed resources? This statement may even be misleading, because it is not clear whether these resources refer to the involved brain structures, or to processes that account for the interaction / communication between these structures. Would it not be plausible to assume that the whole (the emergence of WM) is greater than the sum of its parts (structure and process)?

I hope the aforementioned points clarify my question: Should it bother us that the somewhat ambiguous resource concept is so popular in WM research?

Cheers

Tobi

Recent publication that suggests working memory is widely distributed throughout the neocortex, which reinforces Fuster's ideas on this.  The executive brain is played down and seen as part of a complex web involving a sensory to abstract gradient with multiple feedforward and feedback systems that are activated according to task demands. Also, interesting discussion of the similarity of working memory processes to a Turing Machine.

The paper is:  Christophel et al 2017. The Distributed Nature of Working Memory.  Trends in Cognitive Sciences, February 2017,Vol. 21, No 2.

Link:  http://dx.doi.org/10.1016/j.tics.2016.12.007

Dear Derek,

Thanks for your comment.  The article you cite is a very impressive review of the neural aspects of working memory (WM).  Unfortunately, I don’t think it gives us all the credit my colleagues and I deserve.

            1.  The title of the paper, “The Distributed Nature of Working Memory”, has been my thesis for more than 20 years.  True, the authors make several tangential references to an article of mine (Trends Neurosci., 20, 1997), where that thesis passed the first peer review.  But they neglect to mention several of  our more relevant and recent papers:

J.M. Fuster - Distributed memory for both short and long term.  Neurobiology of Learning and Memory. 70:268-274, 1998.

J.M, Fuster, M. Bodner and J.K. Kroger - Cross-modal and cross-temporal association in neurons of frontal cortex.  Nature, 405:347-351, 2000.

J.M. Fuster – The Cognit: A network model of cortical representation.  International Journal of Psychophysiology, 60:125-132, 2006.

Y.D. Zhou, A. Ardestani, and J.M. Fuster – Distributed and associative working memory. Cerebral Cortex, 17:77-i87, 2007.

J.M. Fuster - Cortex and memory: emergence of a new paradigm. Journal of Cognitive Neuroscience, 21:2047-2072, 2009.

J.M. Fuster and S.L. Bressler – Cognit activation: a mechanism enabling temporal integration in working memory.  Trends in Cognitive Sciences, 16:207-218, 2012.

A. Ardestani, W. Shen, F. Darvas, A.W. Toga and J.M. Fuster - Modulation of Frontoparietal Neurovascular Dynamics in Working Memory. J. Cogni. Neuroscience, (1-23 electronic), 2015.

            2.  The authors also neglect to mention the corollaries of the distributed WM notion. These corollaries are now well supported by microelectrode, cryogenic, field-potential, neuroimaging  and neuropsychological data extensively reviewed in our publications:

(a) The network(s) activated in WM has (have) been formed, prior to testing,  by synaptic modulatation of distributed cortical neuronal groups representing the sensory and motor aspects of an experience and its context (e.g., learning a WM task). 

 (b) These networks or cognits are, therefore, items of established (“long-term”) memory activated and updated for time and context.

(c) These networks (cognits) are the basic elements of cognition and therefore are orderly recruited in all cognitive functions (attention, perception, memory, language, intelligence).

(d)  Whereas we were the first to discover the persistent neuronal activity in the prefrontal cortex during WM, we never claimed that this cortex was the site of WM; we do claim however, that the prefrontal cortex and its executive networks play a literally decisive role in WM, as the neural mediators of cross-temporal contingencies at the top of the perception-action cycle.  (J.M. Fuster – Prefrontal Cortex in Decision-Making: The Perception-Action Cycle. In: Decision Neuroscience: An Integrative Perspective. (Dreher and Tremblay, eds.) Academic Press, London pp. 95–105, 2017.

Cheers,

Joaquín         

Dear Joaquin,

A new study on human rarely recorded neurons from medial prefrontal cortex and they observed persistent activity, which may be important in representation of WM contents, although it's not stimulus-specific. 

Persistently active neurons in human medial frontal and medial temporal lobe support working memory, (2017) Jan Kaminski, Shannon Sullivan, Jeffrey M Chung, Ian B Ross, Adam N Mamelak & Ueli Rutishauser.

And they are from UCLA. :)

Yixuan

Thank you, Yixuan. Indeed it may be.  Joaquín 

Joaquin,

As I was reading Christophel et al, I too was somewhat perplexed by the fact they did not give more credit to your extensive research on WM and the executive brain. Perhaps your intervention on ResearchGate might help to redress the obvious shortcoming.  

Derek

Dear Joaquin

I'm the first author of 'The Distributed Nature of Working Memory'. I'm thankful for your interest in our review.

1.) I am going to straight out agree that our review doesn't give enough justice to the history of distributed conceptions of working memory. The brief format of TICS, citations limits and editorial guidelines discouraged us from going deeply into the development of these ideas and their many predecessors and variants (e.g. in human neuropsychology, behavioral work, modeling, electrophysiology and later imaging). I'd say that we could learn a lot from an article that asks how the different views regarding WM storage came about. For this paper, however, our main goal was to review the experimental literature in a more extensive fashion then has been done before.

In our defense, I'd want to mention that we cite what we believe to be the original reference for the conception of distributed WM (Fuster, 1995) and added your paper that followed in 1997 (which might be more accessible online). On a personal note: While we were writing the review I shared my online reference library with some of my coauthors using the password 'fusterwasright'. This is to say: We at no point intended any affront by not citing any particular reference.

2.) I fully agree with (a-c). We omitted discussing these (and many other) points for the sake of brevity. Concerning (d), I tentatively disagree: While the representational properties that we believe to be situated in PFC (e.g. semantic information, high-level spatial maps, prospective representations and representations of temporal order) certainly give it a distinguished role in WM, we see no evidence that 'decision making' or 'control' can be uniquely attributed to the PFC (if this is your position). That is to say: Our hypothesis is that there can be 'decision' or 'control' processes that do not involve the PFC. We are currently working on testing this hypothesis.

I might add that what we call the 'top' or 'bottom' in the perception-action cycle is a question of perspective: It is clear that in the highly complex grid-structure that we call 'the cortex', activity in any individual area is likely able to affect activity in any other area directly or indirectly. Given that these areas have different representational properties, the question whether they are currently 'in control' or 'being controlled' depends on the moment-to-moment importance of their representational properties.

Many thanks for your interest in our work.

Cheers

Thomas

P.S.: Thanks to Sebo Uithol for bringing this thread to my attention

Dear Thomas:

            Many thanks for your gracious reply.  Some humor is here in order.  On several occasions in the course of my research career I have experienced a fateful sequence of opinions about my findings or ideas:  “It’s not true….; it’s not new…; it’s not you…!”  After reading your paper, I must confess, I felt that much of my work had yet to pass stage two.  Now I feel much better about it, especially after learning that someone--bless his heart—has used my credibility as a password!  I even sense I may not have to die (a necessary condition according to your namesake Kuhn) for my cognit paradigm to be accepted!

            It seems that we agree on most everything.  Now on point (d), where there seems to be some discrepancy between us, I have no problem with the presence in PFC of networks that, as you say, represent semantic information, high-level spatial maps and temporal order.  I would simply qualify those representations as relating, in one way or another, to action.  From my perspective, all of that consists of relatively abstract gestalts of action, which I place under the rubric of high-level executive memory.  Some of that memory has undoubtedly what you term “prospective” attributes; it is “memory of the future” (Ingvar), which, of course, includes plans of action.

            In fact, I go further.  I argue that all executive cognitive functions are essentially “prospective.”  This is true for planning (by definition), attentive set, and even WM, the distributed function at the center of our discussion.  Keep in mind that, within  the perception-action (PA) cycle, WM has two temporal perspectives: past and future.  We seem to forget too easily that, as Baddeley defined it, WM is the retention of information for a prospective action:  the performance of a task or the solution of a problem.

            And the same prospective aspect applies to decision-making.  Here is where we differ; but do we?  If we consider that decisions can get started, “ignited” (Braitenberg), anywhere in the PA cycle, in the brain or in the environment; if we further consider that the information to be retained or decided upon is distributed all over the cortex; and if we further consider that most of our daily-life decisions are automatic and unconscious; then the localization of decision-making in the PFC becomes untenable. In fact, the cortex may not intervene at all in many decisions, and the prefrontal executive “homunculus” falls down to pieces. 

            Here let me share with you a bit of past history that now sounds almost ludicrous:  At some point somebody found “command” neurons (for eye movement) in parietal cortex.  Others found them in the prefrontal cortex (FEF), and an acrimonious dispute ensued about the location of a “command center.”  Until somebody figured out there was no such center anywhere:  parietal and prefrontal neurons were linked by recurrent circuits (PA cycles) and worked jointly in the control of eye movements.

            But now, if we consider that important decisions in our life are novel and complex,  and have to deal with ambiguities, uncertainties, risk, and Bayesian choice, then we have to invoke the PFC at the top of the PA cycle; of course, a PFC not making the decision by itself, but in cooperation with other regions of perceptual and executive representation that join it in attention, in WM and in other prospective functions leading to a choice that is necessarily Bayesian, based on prior history, updated hypothesis and probability.  

            Ultimately, in my opinion, decision and action are the outcome of the associative flow of information between and within active cognitive networks (cognits) distributed throughout the cortex and engaged  in the PA cycle.  Temporally and spatially, that flow results from the context-dependent self-organized activation of that widely distributed system of representation, not from a mythical central executive.  The “executive” functions of the PFC derive exclusively from the fact that, within that system, the PFC is a crucial hub of representation for actions--past, present, and future--with high hierarchical ranking, To the best of my knowledge, this is the answer to the rhetorical question I posed at the beginning of this thread.  To conclude, by paraphrasing my Neuroscience of Freedom and Creativity (Cambridge, 1013), I would like to say that what decides, remembers and makes us free is the entirety of the cerebral cortex, especially, when needed, the PFC, which is  the vanguard of evolution in the human brain.

            Thanks and congratulations to you and your collaborators for a splendid contribution.  If my comments sound in any way arrogant, condescending or dogmatic, I assure you they are not intended to be.  Only with caution and humility can we approach such complicated matters. Please keep me apprised of the progress of your research.

 With best wishes,  Joaquín

P.S.  If you have a chance, convey my warm regards to my other friends at the Charité, Max Straschill and Christopher Behrens.

I am interested in whether 'networks' which I assume represent the working of both the conscious and unconscious cognition are genetic-neurologically predetermined or are they to a larger extent epigenetically fashioned?

Dear Edward:

            Some networks are innate, genetically transmitted, mainly in the phylogenetically and ontogenetically older parts of the nervous system, including the primary sensory and motor areas of the cerebral cortex.  Other networks are formed epigenetically, with the experience of the organism.  We know that learning and interactions with the environment make new networks or expand old ones.  There are several neural processes that lead to that:  sprouting of dendritic branches, synaptic modulation, formation of non-synaptic contacts and others.  The ratio of genetic vs. epigenetic networks varies with the region, the age of the individual and the richness of its life experience.  But there are some general principles that apply to the cortex.  In the cortical phenotype, network epigenesis is more prevalent in associative areas than in primary areas.  And it is a plausible hypothesis that it augments with age and experience.  There are two important countercurrents that we must reckon with, however: (a) at the beginning of life there is considerable attrition of neurons and connections that goes along with the specialization of functions, and (b) with advancing age, there is gradual attrition due to the aging of those neural elements.  You can see how complex are the vicissitudes of cognitive networks in the normal brain.

Cheers'

Joaquín

Dear All,

Here is my chapter on Decision-making.  Joaquin

Thank you Joaquin, I am in full agreement with respect to the associative networks which I find to be six times more active in the left cerebral hemisphere.