Exploring Neural Responses to Stimuli: Mapping Behavior in Simple Life Forms

[Hlafordlaes]

Hello,

I am a layman in biology with a good notion of method only. My background is in linguistics (don't run away yet), but my time has been spent over the last several years honing my understanding in a variety of fields. (Fool with time on his hands syndrome.)

I wish to answer for myself some questions regarding current research about how information is represented in biology. I am most interested in any work relating to tracking molecular and other responses in simple neural systems in response to stimuli.

I have intuitions but not fixed ideas, so no personal agenda to impose on the field from woo land. (I am not interested in QM or new age ideas.)

Eventually, what I would hope to examine (in the research of others, I am afraid) is the possibility for observing any sequential replay of neuron or chemical states in response to a given stimulus. Ideally, this would relate to the set of responses to that class of stimuli, and figure as a prior observed activity to any actual response to that stimulus by the organism. That's my naive concept for the act of choice; I expect to have it contradicted forthwith by reality.

Work mapping responses in simple neural networks of lower life forms is what I'd like to read up on. If there are comparisons of behavior in relation to comparisons among similar networks as so mapped/observed, even better (structural enablement of behavior as a result of blind evolution).

I realize that I have a tremendous distance to travel before knowledgeably handling any answers you may provide. I would welcome direct discouragement if I am grotesquely off track, and redirection to whatever you might find a more appropriate approach to getting a better sense of how information is coded in biology, and its relation to behavior in simple or complex life forms.

I realize I am inquiring about a very difficult area, but once again, rest assured, I am open to any suggestions from professionals in the field, and have no agenda. I have a budget of four to six months to dedicate to reading up, so am not expecting instant miracles, nor to be able to travel very far.

Thank you in advance for any insights you might provide.

Hlaf.

 

[Mike H]

I can't help you with anything neural in nature - never worked with anything more than unicellular organisms (and some eukaryotic cell cultures on occasion).

However, if you're interested in information processing in biology at the simplest possible level - which it sounds like you are - then I'd have to suggest the possibility of looking at bacterial chemotaxis. (Apologies for the Wikipedia link, but it's just to get you started.) It's probably the simplest and best-studied information processing system in biology - in short, how do bacteria decide to change their motion in the search of food? After all, they're unicellular organisms, so there's no 'nervous system' at play here - it's all interlocking cycles of chemistry within the bacterial cell.

I would be happy to provide more references/reading material upon request, as well as answer questions on the topic.

 

[Hlafordlaes]

Mike H:

Perfect! Exactly the kind of thing I was looking for. I think it is the right place for me to get my feet wet first.

I am hoping to apply the same structuralist methods based on behavior and hard data one can do in linguistics without getting lost in "meaning," if only to get a sense of where things are for my own. I'll soon see just how silly that might prove in practice.

Gosh would I love a data set stepping through increasingly complex neural systems with comparative behavior, for the kind of information-agnostic analysis one might do on that basis. I am quite sure a microbiologist would tell me "hold on, you need to look at the cell level first, it changes before other structures." Then the biochemist will tell me I'd need to drill down to protein folding to witness any action, and a cellular view is not granular enough.

I just hope the turtles stop somewhere, before that pesky IQ limit of mine kicks in. Thanks again!

 

[Mike H]

If the focus is in understanding the 'logical circuitry' of information processing here - namely, how the bacterial cell detects signals, processes them, and then adjusts its behavior at a logical (if not molecular & mechanistic) level - then I suspect the current state of knowledge will probably serve you well. While a number of detailed mechanistic details are still unresolved (and will likely continue to remain so for some time), it hasn't served as a severe impediment for such analyses by many over the years.

I have no real clue what you are alluding to when you mention 'structuralist methods' as I am poorly read in linguistics, so I can't provide any useful inter/transdisciplinary commentary there. If you're plotting on trying to move onto multicellular organisms where the nervous system has been quite well-characterized, my recollection is that there are some where neural development has been traced from fertilization through the embryonic stage (worms, zebrafish, Drosophila) to various extents. You'll want to check my memory on that, though. I know we have some neuroscientists floating around on here with a theoretical bent, so perhaps they can offer some useful words of advice.

 

[Hlafordlaes]

Background
Like many, my curiosity started at the far other end of the discussions regarding mind and brain, and to mention a taboo word softly and with trepidation: consciousness. Those conversations are filled with speculation and rampant rationalist fantasy. I've had quite enough of that, thanks!

I would rather inform my views with whatever data is available, and not fill in any gaps with funny stuff.

Structural linguistics was a good response to the problems of dealing with meaning when examining code, in order to avoid just that, rationalist fantasies. Essentially, the answer to the problem of meaning was to ignore it almost entirely. It is no surprise that the method basically relies on holding all but a very small element constant and comparing meanings between two structures that are minimally different. ("Meaning" here is simply a question a native speaker stating "I agree those two samples are different." Dictionary meanings are ignored.)

Generative syntax (think Chomsky) was an important advance in positing underlying code in terms of purported intermediate cognitive steps between thought ("don't go there") and surface expression, or utterance ("grammar"). Some of that has been born out experimentally by studying rule deletion and modification due to genetic disease or brain trauma. (Linguists are ambulance chasers, in that sense). Below all that, we have folks working directly on neural structures related to language, "underneath" meaning.

So, in spite of ongoing research into semantics, the field lives on based on a foundation in neurology, a ground floor of messy meaning that still sort of sits apart in philosophy departments, and 'upper floors' consisting of an observable coding process amenable to scientific method. IOW, a structuralist approach to simple organisms in my mind is one that focuses on extremely small differences in physical configuration with observable behavior, avoiding all along any mysticism about what "thinking" is.

[To make matters worse, I am a former linguist (had to leave the field some decades ago since the regime I came to be living in prohibited foreigners from research or univ. teaching positions), and am old now. Depending the day and time of posting, I can range from decent clarity to thick London fog.]

...

Translating the above to my goals, stated naively from outside biology:

Examine the earliest mechanisms, cellular or in the form of rudimentary nervous systems, to track where behavior at a logical level can be seen to change as a function of changes in those mechanisms. (So I am delighted to think there is a good current state of knowledge in this regard, and want to read all about it.)

Ideally, at some point we have the earliest logical circuitry that we can see shows preferential decision making. That is, to take a spin on chemotaxis as if that was all we needed to take into account: in the presence of competing chemoattractants, in what way are they compared? How does one chemoattractant "win" over another for the organisms attention?

The "biggies" for me: When can it be said that an organism is sophisticated enough to make choices based on the valences of historical outcomes? When are choices no longer a comparison of gradients, but weighted outcomes based on history? What physical circuitry is required?

My mere intuitions indicate that at some level, any organism we recognize as "decision making agent" must match an external stimulus to an internal template for it that yields a valence allowing choice. Is that a sequential comparison of molecular representations checking for optimal choice, or something else? Is simple memory a map of chemoattractors/repellants to the valences of historical outcomes?

To go further in my madness: Does a successful purported "replay" of the internal template or molecular representation also cause a cascade of simulated responses in the simple nervous system? That is, is experience nearly fully remodeled internally in some chemical fashion? (Certainly in humans this is the subjective case for some but not all stimuli.)

The long-term, highly provisional, woo-endangered thing I posit has to do with thinking of the qualia of consciousness ~ those bugaboos that seem to lull everyone into mysticism ~ as merely vibrant systemic responses of replay. Before I can get anywhere near to expressing that in a responsible way, to be respectful of science I'll need to start at wherever the ground floor is, and spend what's left of my life working my way up.

What I mean, then, by taking a structuralist approach, is to in this case avoid positing symbolic values to molecular representations, and only look at relations between logical systems and behavior. Some day, though, one would hope we might attach some sort of "meaning" or external objective valence to a particular molecule or chain (physical representation), but as an ex-linguist, that is the area I already expect to be quite, quite difficult, if ever possible.

I wish to scrutinize the literature on simple systems to see how far we may have identified the relation between incremental steps in behavior and small changes in cell or system. I am obviously now at the stage of asking beginner questions, and narrowing the proper areas of research to look at. I possesses all the shortcomings of age and amateur status.

 

[Mike H]

Trying to digest all of the above might be a bit much for me as the work week picks up, but let me try to respond to some questions that are fairly straightforward (at least to me at first pass).

The bacterium seems to integrate all available data to it in terms of compounds detected (there are, in E. coli at least, five different type of chemosensing proteins, and each can detect a couple of attractants and repellents), and not only from the chemosensing array embedded in the membrane - recent work has shown that the ability of bacteria to take up glucose also gets included in the 'decision-making.' Having said that, there's been some interesting results that suggest that it's the number of particular receptor types that might be governing the response in the presence of multiple attractants, which can vary with the growth conditions. Much older work looking at behavior in the presence of an attractant and repellent show that if one holds the concentration of repellent steady, one can add in enough attractant to get the bacterium to move up the combined gradient of attractant and repellent, but if one adds too much attractant, one will saturate the sensory mechanism and the bacterium will again move down the gradient away from the attractant and repellent.

We do know that bacteria use a temporal mechanism for sensing - there's been a fair amount of analysis to understanding physical bounds for spatial and temporal sensing of stimuli by (micro)organisms and the locomotion of microorganisms. The small size of bacteria, relative to eukaryotic cells, make spatially-based sensing (e.g., detecting a gradient along its length) problematic. There's a feedback loop that essentially let's the bacterium compare its present chemical environment to one from a few seconds ago to decide whether to change its behavior (which, for bacteria, is easy to observe - keep swimming in the same direction or tumble to a new direction). It's a short-term 'memory' - there's no indication (to date, at least) that it retains any information of its state from, say, an hour ago.

Here's a recent review that might be helpful as well to look at, especially as the two authors are actually involved in many of these studies.

 

[Hlafordlaes]

Mike H,

Thank you so very much for putting that together. I think you understand perfectly the approach I would like to take. I think my next step is to digest more of what your two helpful posts have provided; that should take quite some time.

The onus is now on me to get into high gear. I'll be joining (as affiliate) the Society for Neuroscience, plus attempting to gain any third-party associations I might to those places seeming richest in data and closest to my interests, followed by grabbing some textbooks off Amazon and so forth.

Today I was thinking that my first "mecca" will be identifying the simplest organism whose behavior seems to indicate some rudimentary form of retaining learned reactions, and the info you provided certainly helps me get the background to begin to look at this kind of issue.

Thanks again for everything,
Hlaf.

 

[Hornbein]

There is a lot of research in this area. Bacterial chemotaxis is very basic (that's good). As the next step you might like Nobel Prize winner in neurobiology Eric Kandel's book about his study of the nervous systems of giant snails. It includes autobiographical info about Nazi Austria, which was even more anti-Semitic than Germany.

 

[Mike H]

As I've already pointed out NIH's PubMedCentral (which serves as an open-access repository for federally funded research) with some of my links, I should also probably mention the Bookshelf at NIH, which might find to be of interest, especially in terms of possibly cutting down on total Amazon expenditures, depending on what you were planning on reading. (At the very least, it may help save some time searching for so-called 'textbook' knowledge that might be presumed at various points along the way.)

I would second Hornbein's suggestion to check out the work done on Aplysia californica, especially if you're interested in organisms capable of learning. It's also been a model system for other biological research to some extent, so that should help provide some extra appropriate background reading as you find necessary/illuminating.

 

[Hlafordlaes]

Gosh, thanks again everyone. I think I am about to fall in love with a snail. And a bookshelf, too.

I was going to publish a joke or two in the Fun section as my only means of providing meager compensation, but the ones I wanted to use have been posted once or twice elsewhere, so for SEO reasons I cannot republish them here. Don't want to link to things off board, so no luck.

 

[Pythagorean]

With regard to neural networks, you might be interested in "functional motifs". Am on mobile now so just one link:

http://journals.plos.org/plosbiology/article?id=10.1371/journal.pbio.0020369#pbio-0020369-g005

 

[Hlafordlaes]

[@Pythagorean: Imagine Spock lost in the viewer looking at a new thing, and repeating "Fascinating!" My hope is that maybe in a year or two, I'll be ready to appreciate the content more and close my jaw. For now, I am still seeking out the simple. To wit:]

The Curious Incident of the Infinitely Receding Problem Statement​

I am sure you have faced this yourself. You start out with a nagging feeling there is a gap in your knowledge, and set out to solve the problem. You find the original problem has been solved. A new one arises, yet once again, it has been solved. Eager to push a frontier in the exciting realm you have entered, you furiously reformulate and forge on. Rinse and repeat, and you have the grad student thesis nightmare (or a new devotee of string theory /jk).

My original problem was (is!*) to identify early nervous systems with simple behaviors and get a handle on linguistics's dark cousin: intelligence. Specifically, that part dealing with physical reference systems, not symbolic systems, such as human language. I have found (was kindly led to) a path paved from 1881 (chemotaxis) until a recent Nobel prize involving a snail, answering a long string of pet problem statements. Even if one secretly knows from past experience that this, not can, but will happen with problem statements, it simultaneously is a source of muttered praises and curses (well, in my cantankerous case, anyway).

While I of course need months/years to absorb and meaningfully learn the path already laid out, I am going to venture a new problem statement, in the odd hope that in spite of doing the same thing (positing in a relative vacuum), I may get a different result (not something entirely old). Definition of madness, iirc, but I like having a pet problem statement in my pocket as I read.

It occurs to me that a physical reference system, for now defined as anything an organism uses actively to respond to stimuli using memory, faces issues that a symbolic system does not. That is, a symbolic system can be entirely arbitrary, such as assigning different symbols for the same thing (house|casa). In the case of simple organic environmental interactions, the use of memory as part of response would seem to indicate that the reference cannot be arbitrary.

In other words, if the use of a primitive memory system were arbitrary in reference, we then have, at some stage of neural sophistication beyond the level of chemotaxis, the need for each organism to "understand" which memory reference is valid to select as matching the external stimuli. That would introduce an index problem, or the chemical version of a homunculus. The selection must be automatic. How?

To state as a problem to solve:
In what sense can neural structures and in particular, the molecular or other elements involved be considered, while structurally arbitrary in the sense of being mere evolutionary artifacts, determined in their reference once incorporated into the genome? Alternately, and much less woo-laden, how do relatively simple organisms solve the index problem, and access memories stored in an entirely arbitrary fashion by each individual, if this is the case? ...Or, is there no index problem, boo-hoo?



*Still is a huge problem to solve for me, but seems to be one whose answers I'll need to study and learn, not generate. The above are the thoughts I am having as I happily read along. Please do not take them as a poor reflection on the excellent posters who have responded in thread. My bad, entirely.

 

[Hlafordlaes]

Much quicker question this time: good math books for science majors. Can you recommend one/some?

After my naive forays above, it is now clear the field of interest to me is computational biology.

Luckily, there is a lot of research in that field locally, and I may even be moving soon right next to campus by strange and very timely coincidence. I plan to audit courses or simply do them in parallel privately. Meanwhile, I'll be joining a few professional assns to track what's going on.

I cannot foresee all the different areas in maths I'll need to cover, but it's a done deal I'll need to go back and redo a lot of simple algebra to get my equation-solving back on solid footing, some trig and lots of calculus, plus on the side, statistics.

I am a b¡t wary of the Schaum (?) series ever since a prof decades ago showed me a list of wrong answers in their stats book that he had been sending them corrections for for years, with no changes ever made in subsequent re-editions. Since this is self-study, I do not want to be thrown for unnecessary loops.

Got any good books in mind for science undergrads or dummie new grad students who need math? Better if correct answers are in the back or whatever. Thx.

 

[Mike H]

You might want to go look around (and ask if you can't find anything after searching) in the Science & Math Textbooks subsection of the Academic Guidance section for recommendations, if you haven't already - usually the mathematics teachers/faculty who contribute there will have a better idea of what is out there and each text's respective strengths & weaknesses, especially for pedagogical purposes. I'm not just how many mathematics texts are written for scientists at the levels you're asking about (that is, at a algebra/precalculus to calculus level) - what I've seen has been written for undergraduates who have had at least a year's worth of single-variable calculus in my observations (e.g., something like Boas' "Mathematical Methods in the Physical Sciences), if not more preparation.

You'll just have to be patient wading through "computational biology" literature - it can range from everything from ecology to biomolecular simulation, depending on who's using the term. My experience is that people will usually try to specify their general field of interest (e.g., computational neuroscience vs computational genomics vs. ... and so on).

 

[Pythagorean]

Alternately, and much less woo-laden, how do relatively simple organisms solve the index problem, and access memories stored in an entirely arbitrary fashion by each individual, if this is the case?

Well, genetics itself is kind of an indexing system isn't it (depending on what you mean). DNA methylation (a mechanism of epigenetics) seem like it could be considered and indexing system.

It's a controversial issue whether molecular networks work similarly in an organism. There are lots of models and ideas proposed, but no clear evidence. For example:

We demonstrate how a single-celled organism could undertake associative learning. Although to date only one previous study has found experimental evidence for such learning, there is no reason in principle why it should not occur. We propose a gene regulatory network that is capable of associative learning between any pre-specified set of chemical signals, in a Hebbian manner, within a single cell. A mathematical model is developed, and simulations show a clear learned response. A preliminary design for implementing this model using plasmids within Escherichia coli is presented, along with an alternative approach, based on double-phosphorylated protein kinases.

http://rsif.royalsocietypublishing.org/content/6/34/463.long

 

[Hlafordlaes]

Thanks, Mike H and Pythagorean. With luck, I'll be following coursework at the Institut de Neurociencies at the Univ. of Barcelona in the Fall, either auditing or enrolled. I'll spend all summer purely on maths. Some good conferences are also on schedule locally in September.

As Mike H points out, I am finding this is not really well-defined field, as there are so many ways of approaching topics around brains and what they do. In the end, you just have to dive in and then see where things take you. But that is also the plus side: newcomers may find a fresh niche or two. Between this field and neurolinguistics, I am sure there is a layer somewhere I can work on.

Pythagorean, fascinating that link to assoc. learning in a single cell. My wild guess is that there ought to be quite a lot of, for lack of a better word, "thinking" at even the simple end of the spectrum.

Well, at this point I have my math books all selected, and just have to wait a month or so to get into my new digs near the university. I feel like a kid again!