Brain Physiology, Cognition and Consciousness group: topic
This is a public discussion board
From Memory System to Mind, How Mind Might Work
Graeme Smith
Thursday, 16 July 2009 21:15 UTC
This is a companion to From Neuron to Memory System: How Memory Might Work Questions about memory should be fielded on that topic, rather than this one, this one is about mind, or as some have taken to calling it Consciousness.
There has been a lot of discussion about consciousness here in BPCC, and I admit I haven’t kept abreast of it, if only because I disagree with a lot that has been said. I am hoping to present in this discussion topic my ideas around consciousness. I expect that I will be stepping on some toes when I do.
My Philosophical Viewpoint is simple and I have held it for a number of years so I don’t expect it will change. Consciousness is probably something that can be explained physically, and we shouldn’t look for a non-physical solution until we have explained enough about the rest of the brains function to eliminate consciousness from the brains normal function.
We haven’t, so I don’t look for a non-physical or even exotic physical solution.
Further, I believe that it is possible to create a machine that is conscious
because once we have explained most of the brain function, we will have a blue-print for a conscious machine.
However I do not believe that we have conscious machines yet, if only because we have not yet designed one that is based on a sufficiently sophisticated model of the brain to meet my ideas of consciousness.
What happened a few years back, was I went from feeling that this was true, to feeling that I had an idea of how we could build such a machine, it was an intuitive jump, that I have been years trying to prove was actually possible.
My earliest theories were easy to shoot down, but as I have begun to study the brain, it has become harder and harder to convince me that my ideas are wrong. They are however to some extent novel, and coming from nowhere as I am, I expect more than a modicum of resistance to them.
The basic intuitive jump is still valid, what has changed is the architecture by which I try to achieve it. Today, my theory involves a basic cognitive architecture based on a hybrid memory model, and a Virtual Machine that sits on top of that architecture, to actually implement consciousness.
The Mind Model begins of course with my memory model which in turn is based on a model of how cortex tissues work. Mature Isocortical Tissue is seen as a three laminae implicit memory, and a two layer interface that allows it to be explicitly accessed.
Of importance to the memory model is the constraints on implicit memory, and how they limit the implementation of an explicit memory complicating the interface. As a result of this, and other parallel memory loops in the larger system, I trace somewhere around 9 separate forms or phases of attention.
The first level of attention, the quick response network, is instinctively driven and operates at about the level of a conditioned reflex. The second level of attention, the Functional Cluster level, is instinctively driven again, and consists of selecting zones of salience for later processing.
The zones are individually prioritized and the zones with the greatest salience are then converted into chunks. Automatically by the Bottleneck.
Making sense of the chunks however, is not a pre-programmed function and so we should see in the fourth and later attention stages some sort of control progression being implemented.
The progression I have suggested, is the Random Impulse, Intention, Volition progression. Random Impulse based motivation depends on limbic bias, to steer it and this starts out to be instinctively steered in the naive state.
As a library of sequences of action/processing begins to form in the cerebellum, it becomes possible to pre-activate sequences according to context. As more sequences are stored, it becomes necessary to select from among many parallel sequences. The Intention System which controls the selection of which sequence is best for which novel situation, is born. However in some cases the selection from among pre-recorded sequences is not sufficient, and so a system that can create programs to order is needed, and we call that system consciousness, and the control system it creates volition.
Now it is important to realize that this model of the control systems, is based on the function of the ACC, which selects from among multiple targets which pre-activated sequences will be allowed to complete. So Intention and Volition are actually responses to the problems of the ACC resolving a problem in selection.
This means that it is probably the circuitry around the ACC in the prefrontal cortex that controls whether we experience Intention or Volition.
Those that see Consciousness and Volition as being the primary cause of activity, will not be willing to accept this model.
Now I am going to shift perspective on this model from control, to programming. Essentially any sequence of action/processing commands, is a program in the command language. A sequence of commands if inserted as a lump, into the stream of commands, can act as a macro command. All that is needed is a two-loop interpreter. In this model we see the inner loop as being achieved by the cerebellum, and the outer loop being achieved by the SMA.
Intention is a method of selecting from among macros, a stream of macros that meets the limbic biases of the individual. Part of this process is the evaluation of a macro to see if it will invalidate the comfort zone of the individual. Part of this process is the creation of a log, that allows the system to automatically rewind a program that has gone astray. To do that we need to have a listing of all the processes started during intention. Part of the problem is that we may have multiple processes ongoing at the same time, and so there might be a competition for place on the more or less serial log. I have called this log the “Awareness” buffer because the processing of the log in order to determine whether or not to rewind, makes us aware that we were processing information, and Echoes that information to our awareness.
Awareness is required simply because we allow macros to be formed automatically, and they might exceed the comfort zone of the individual organism and have to be rewound so that we can protect ourselves from extreme situations that were errors in the automatic programming.
Beyond intention lies the rewind script, and the problem of how to select a new macro, that does not have the flaws of the old one. To deal with this, we need to be able to monitor the parameters under which the old macro ran, and have some sort of feedback to the limbic system that allows us to adjust limbic biases. Further we need some way to evaluate, whether we have adjusted the limbic biases correctly, and we need some way of fabricating a new program around the new limbic values. One way of doing this is to have limbic meta-cognitive signals that indicate the strength of the limbic biases, Awareness plus these meta-cognitive signals, gives us a greater number of options for redirecting intention. We can then experiment with different variations of intention, and see how they interact with the model of self.
Of interest here is the “Consciousness” Buffer which reflects the richer feedback environment including the meta-cognitive signals so that second order processes can be rewound as easily as first order processes.
By presenting the meta-cognitive signals as “Feelings” the system assures that the organism adjusts it’s parameters in ways that increase the comfort rather than making it worse, unless there is a valid reason why not. This allows the brain to readjust the “model of self” if it gets out of range.
To be able to do that however, the brain has to recognize which elements it caused itself and which came from the outside environment, and to do this, it uses a meta-cognitive signal associated with the Awareness/intention mechanism to indicate which functions it pre-approved. One of the ways that we might be able to tell if we are experiencing awareness or consciousness is by testing whether or not we are aware of a feeling of self. If the feeling exists we are probably experiencing consciousness rather than awareness.
-
Replies
Jump to resultsResults
-
“Well chris, IDM is your baby and I understand your adherence to it. but really, using AM/FM technology as a metaphor gets a little thin.”
? metaphor? I think you are missing some basics re neuron dynamics. The frequencies are spatial and the AM/FM dynamics well covered – A diagram for you
The pulse/wave dynamic, specialist forms of expressing differentiating/integrating, sets the ground for information processing and feedback loops introduce recursion (to implement the exclusive OR in the neurology requires the feedback dynamic of one neuron’s output feeding back into an upstream cell’s input. This dynamic is the basic form of memory)
The pulse/wave dynamic applies all the way ‘up’ from the single neuron to the play of hemispheres of the neocortex (to see such in the form of models of the brain as a low/high band filter see research in such as:
Ivry, R.B., & Robertson, L.C.,(1998) “The Two Sides of Perception” MITP
Hutcheon, B., & Yarom, Y., (2000) “Resonance, oscillation and the intrinsic frequency preferences of neurons” Trends Neurosci. (2000) 23, 216-222
Qiu FT, von der Heydt R.(2005)“Figure and ground in the visual cortex: v2 combines stereoscopic cues with gestalt rules” Neuron. 2005 Jul 7;47(1):155-66.
Heinen K, Jolij J, Lamme VA.(2005)“Figure-ground segregation requires two distinct periods of activity in V1: a transcranial magnetic stimulation study” Neuroreport. 2005 Sep 8;16(13):1483-1487
Keri S, Kelemen O, Benedek G, Janka Z.(2005)“Lateral interactions in the visual cortex of patients with schizophrenia and bipolar disorder.” Psychol Med. 2005 Jul;35(7):1043-51
Nakayama K.(2005)“Resolving border disputes in midlevel vision” Neuron. 2005 Jul 7;47(1):5-8
van der Togt C, Kalitzin S, Spekreijse H, Lamme VA, Super H.(2005)“Synchrony dynamics in monkey V1 predict success in visual detection.” Cereb Cortex. 2005 Apr 20
The mixing of specialist sense data reflects the conversion of sense data to patterns of frequencies/wavelengths/amplitudes. Such mixing allows for the experience of synaesthesia etc and the constancy of classes of meaning indicate such grounded in the neurology (sense of wholeness, partness etc are universals, to what they are applied is local context stuff. For coverage on combination of sense data see such as:
Stein, B.E., and Meredith, M.A., (1993) “The Merging of the Senses” MITP or see my page on synaesthesia issues )
“And that combinational stuff, you keep pushing, seems rather contrived when you consider that the locations of signals is indeterminate at each level of organization in the brain. The problem is that it is so easy to see patterns in processes, because we as humans expect to see patterns in processes.”
Which is the point. Our neurology is structured to deal with pattern matching and the set of POSSIBLE patterns is identified in the IDM work. The dynamics of networks introduce the ‘small world’ network development from the set of what is possible and as we learn more and more so the small world is refined into an actualised regular network (as compared to the originating regular network that covers potentials). Or more so the set of specialist smaller worlds that develop from the small world sum to give us details of actualisations of all possibles.
ANY attempt to map the brain and on to dealing with describing “Consciousness” will utilise the BASE level classes of meanings that come out of the neurology; we cannot avoid this as this level sets the ground for all meaning that is then extended through use of symbols where such allow us to communicate outside of the base level context but using the classes of meanings derived at that level.
Thus, if one understands what is POSSIBLE so one can gain insights into distorted perspectives that have resulted from ad-hoc development/understanding of how meaning is derived. We have the SAME patterns applying at ALL scales of analysis since our neurology works as a filtering system and in doing so determines the structure of all meaning that is sharable. The degree of development at a level reflects differentiation pressures and so influences more abstract levels where THEY require well-differentiated support structures to work.
Trying to map what the brain does merely shows us a LOCAL adaptation in managing patterns of differentiating/integrating, anti-symmetry/symmetry, where such are fundamental to the universe (the etymology of the neuron goes back some 600 million years to sponge life and so reflects the success of the pulse/wave, differentiating/integrating, dynamic as ‘reflective’ of the environment)
Since the brain reflects a considerable time period of ad-hoc development, identifying what that development is trying to do aids in understanding and that includes ability to advance such development through understanding what is possible given the neurology dynamics.
“I wish you could be more direct about what you are trying to say, I find myself overwhelmed with the prospect of trying to tie all the threads of your work together into a cohesive concept.”
This general forum is focused on consciousness with this particular list started by you trying to cover brain-mind dynamics. The IDM work covers all that is POSSIBLE in the dynamics from the meaning perspective; all specialist patterns will be in the form of sets of labels that ‘point’ to the IDM classes of meanings. Included in this will be all POSSIBLE classes of consciousness given the neurology that supports such.
Hans etc are trying to understand consciousness etc from a bottom-up perspective but in doing so each contributor feels THEIR point of view is the ‘best fit’ – the IDM perspective is that ALL of the points of view apply since the description of consciousness will be through the set of POSSIBLE meanings available to each of us as neuron-dependent life forms.
By being able to identify the classes of consciousness gives us a head start in mapping out the details of the classes and on into the instantiation of a class in the form of a unique mindset (where such is still dependent upon our species nature to support such a mindset)
“Recursion is an important factor in learning, but it is not magical, everything that is done by recursion can be done by iterative means, perhaps not as efficiently, but without a doubt.”
No. The emergence of an organic element FROM a mechanistic element is through recursive action. The essential feature in any language is the ability for the language to describe itself. What the IDM work does is identify how this is possible without the precision of words etc where the implicit SYMMETRY present encodes the ‘all is connected’ nature in the classes of meanings derived from recursion and used as a language.
In other words, the INITIAL conditions of recursion are mechanistic and potentially heading into an infinite regress. GIVEN depth in the recursion we find emerge an organic element that serves as a brake to the regression. This feature has NOT been identified in the usual texts covering recursion.
What is being dealt with here is the PARALLEL nature of our being and the ability to process data holistically and convert such to SERIAL forms of information and the reverse. Understanding these features of the neurology take us way beyond the current perspectives presented by the specialists we find in these forums – their focus on the trees has forced them to miss the dynamics of the forest – understandably so since the data leading to identifying the IDM categories etc has only been around for the last few decades or so and is not ‘mainstream’ but the frustrations presented by many of them suggest it is time for a ‘new paradigm’ and IDM can aid in such development ;-)
Chris.
-
Ok look Chris recursion gets it’s power from symmetry but I have to say that it is not magic, like I said there is nothing that you can do with recursion that you could not also do iteratively. I know this because when I was training in Computer Science, we looked into the nature of the binary search, and my instructor picked the recursion apart for us, and showed that it was a basically iterative approach that took advantage of the symmetry of the search process to reduce the number of iterations required.
In computing a recursive process is a process that calls itself. The binary search, works by choosing a leaf in the binary representation, and feeding that leaf the rest of the search terms. The reason it works so well, is simply that every decision can be reduced to deciding if the search term is greater or lesser than the value of the next item in the list. A similar architecture can be used in the binary sort. Most forms of recursion are well symbolized by the recursive nature of the binary search/sort pattern. Yes symmetry speeds the process by reducing the number of elements that have to be searched in any particular path, but no it is not anything more than a more efficient form of an iterative process.
Results
-
