New Research Suggests that Diversity in Dogs Comes from the Wolf's DNA

    • Gold Top Dog

    Hi Ron,

    Thanks for your response.

    First of all,I don't know what you mean when you say "Deciding that vocal praise is the only true reward is human arrogance" What is that in reference to? Did I give you the impression that I thought this was so? This seems a very odd thing to say.

    Secondly, you seem to keep missing my point about why I directed you to Dr. Gallistel's paper.

    1) You made a statement that I was "quoting one guy." 

    2) I directed you to the references section of that paper to show that my reference to Dr. Gallistel's work is not about "quoting one guy," that, in fact, there's a whole body of work, being done by numerous scientists, on this specific problem.

    Do they all agree? Of course not. But the point was, and still is; it's not just one guy. 

    Has the issue been proven conclusively in Gallistel's (and others';) favor? No. But I would argue that the fact that there are these holes in learning theory, holes that are becoming clearer each day, strongly suggests that some of the basic tenets of behavioral science haven't been proved conclusively either. At this point, in the 21st Century, it's become a bit outdated.

    You told a story about how you got Shadow to give up a cotton rat in favor of a treat, and you believe that shows the efficacy of operant conditioning. This may be true, but remember, operant conditioning is based primarily, at its root level, on the behaviors of captive animals, kept in an enclosed, constricted environment: pigeons in boxes pecking levers, rats running through mazes, etc.

    How successful would you have been if you had tried redirecting Shadow from chasing a rat in an open, natural environment? I'm not saying it couldn't be done, I'm just asking, on a scale of one to ten, if the incident in the house were a 9 or 10, where would your ability to influence the dog's behavior in a more natural setting lie? In your view, does treat-in-exchange-for-rat have a higher or lesser value indoors or out?

    ron2
    A dog praised for not scavenging in your presence is not an example of an aversive to stop the behavior of scavenging in your presence.

    First of all, the dog isn't being praised for not scavenging. He's being praised while he's actively engaged in the behavior. And if you do this properly, and often enough (with Freddie it only took me 3 days), you can extinguish the scavenging behavior completely, simply by praising the dog every time he does the very behavior you don't want him to do. (There are other examples of this besides scavenging; barking at the door is one that springs to mind.) And according to the principles of OC, using praise to extinguish the behavior you're praising would mean that in this case, praise is either an aversive stimulus or a form of positive punishment. And frankly, that doesn't make any sense.

    ron2
    It is an example of a dog for whom praise is the highest reward redirecting behavior or extinguishing the behavior of scavenging in your presence in the presence of the greater reward for what you want. OC.

    I'm not clear about what this means. It bears very little resemblance to normal OC terminology, or any that I'm aware of. Perhaps you could clarify.

    However, going back to the dogs-as-scavengers theory, and away from the possible holes that exist in operant conditioning, if survival instincts are paramount, then where do these dogs come from, the ones "for whom praise is the highest reward?" If such dogs exist (and I'm not doubting that they do), then it would mean that the model of canine behavior this thread (or part of it) is based on -- meaning the wolf's ability to override his own survival instincts in favor of acting in a cooperative social manner while hunting -- has had a direct influence on how modern domesticated dogs respond to their owners.

    Anyway, that's how I see it,

    LCK

     

    • Gold Top Dog

    spiritdogs

    Don't get me wrong. OC works much of the time. But it isn't 100% accurate, or 100% effective. And dogs are our clearest window into how and why it doesn't work all the time. To repeat what I've said before, I'm not suggesting throwing the baby out with the bathwater. I'm simply questioning the conventional wisdom behind +R training.

    LCK, I hate to be the bearer of bad news here, but NOTHING is 100% accurate or effective.  

     

    That's funny.

    Of course nothing is 100% accurate or effective. However, in terms of what's being presented in the marketplace -- or what was being presented at the outset of the modern training movement (Dunbar and others are now backing down from their original rhetoric) -- when dog owners hear words and phrases like "scientific" or "based on the science of how animals learn," many of them are expecting something that actually rises to the level of 100% effectiveness.

    LCK

    • Gold Top Dog

    Alright then, LCK, you weren't just quoting Gallistel. You were, by your own definition here, quoting all the other researchers listed in the references. So, every single one of them is on board with Gallistel's desire to "dismantle OC," right? They're all on board and agree to publication of this report about the differences between long-term dopamine and momentary, event-related dopamine and that learning took place and that event-related dopamine existed in both "rewarding" and "non-rewarding" situations. Though there wasn't an exact definition of how they defined what was rewarding and not rewarding for the dog, let's just accept that summation for the moment. So, then, can we assume by your connection to these other references that the other researchers are also on board with your theory? I know that sounds far reaching but you are the one insisting you are also referring to the work of the other scientists referenced therein.

    Shadow will bark at the tv if there is a dog on there. I've told him "you go boy! Go get em" in a good voice, not punishing. Guess what? For some reason, he keeps barking and it doesn't extinguish. If I could teach him to read, maybe he could read NDT and get with the program.Wink

    You don't suppose that when you praise the dog in these circumstances, he turns to you and ignores the other thing because turning to you gives him the greater reward, your praise? For my dog, vocal praise is not reward number one. Food treats are.

    As for saying that that the assumption that vocal praise is the highest reward to a dog is to ignore the simple fact that the dog decides what is rewarding, in a sense. That is,some dogs find one thing more rewarding than any other, not exactly a conscious decision but a fact of life. To decide that what you (in general) thinks is most rewarding to the dog is arrogant, because one is not considering the dog's individual idiosyncracies. I was not directly saying that you are arrogant. Evidently, you have been fortunate to have a dog for whom vocal praise is the mother lode of rewards. For my dog, it's roast beef.

    Shadow was barking at the little kids next door. I went over beside him. Gently touched and rubbed his neck. Slowly redirecting him to me. Called "here" and walked away. And we went in the house. And I gave him a few treats. I'm just bad that way.

    If you are saying that OC didn't work in the prescribed manner, how can you use the term extinguish? A behavior extinguishes because it is no longer rewarding or not as rewarding as something else. And perhaps, your praise has become a cue for redirection that is later rewarded by play, which may be the actual reward.

     

    • Gold Top Dog

    ron2
    Alright then, LCK, you weren't just quoting Gallistel.

    Thank you.

    ron2
    You were, by your own definition here, quoting all the other researchers listed in the references.

    I wasn't quoting anyone. In the blog article I wrote, I was simply using several of Gallistel's papers (three or four of them) as a brush to paint a picture of some of the research going on in this area. I chose that particular paper at random.

    ron2
    So, every single one of them is on board with Gallistel's desire to "dismantle OC," right

    As with most scientific papers or articles, there will be a range of references, from those that support the central thesis, and those that oppose it. So, no. But I had thought we'd gotten beyond that point, Ron. It's still not just me "quoting one guy." It never was. And I'm not sure I ever used the word "dismantle" (though it's a good one because the word mantle conjures images of an emperor, particularly one who isn't wearing any clothes).

    I like Gallistel for a lot of reasons. One of them is his view of read/write memory and the brain.

    ron2
    You don't suppose that when you praise the dog in these circumstances, he turns to you and ignores the other thing because turning to you gives him the greater reward, your praise?

    In the initial part of my experiment, the praise came while the dog had a juicy 1/2 chicken breast in his mouth. The praise caused him to drop it. It was only when I continued to praise him that he turned and came over to me. So, in effect, my praise stopped the scavenging behavior in its tracks, and on subsequent occasions, caused the dog to stop showing any interest in scavenging at all.

    That can't be explained through the principles of behavioral science.

    Other examples come from some of the aversives used to train working dogs, which actually increase the reliability of the behaviors being asked for. For instance, when training search-and-rescue dogs to run up and down fire escapes, some trainers will throw pots and pans at the fire escape while the dogs are going up or down it. This increases their commitment to the behavior. In terms of their ability to navigate rubble, etc., dogs are actually less suitable candidates for working at disaster sites than monkeys or cats. Yet if you threw pots and pans at monkeys and cats, as part of their training, they'd immediately disappear and never come back.

    So there's something unique about dogs that shows us a clear window into some of the flaws of the OC model.

    Also, if praise is the greater reward -- more satisfying than a tasty chicken breast -- then what does that say about the nature of praise? And what does it say about the nature of dogs? (Operant conditioning is used successfully in wolf sanctuaries to enable veterinarians and others to be able to safely handle the animals; could you use praise to get a hungry wolf to drop something he was intent on eating?)

    ron2
    If you are saying that OC didn't work in the prescribed manner, how can you use the term extinguish? A behavior extinguishes because it is no longer rewarding or not as rewarding as something else.

    Right. But reward isn't a technical term in OC. Remember, a reinforcement is anything that, within a temporal relationship, increases the response strength of a certain behavior. A punisher is something that does the opposite. So here praise was acting as a punisher.

    That doesn't make any sense. And yet, it works.

    LCK

    • Gold Top Dog

    The nature of praise is irrelevant. What is important is how the dog responds to it. And reward and punishment are, as noted by many, maybe even yourself, operational terms, and should be free of emotional content for the human. So, you use praise and it stops a behavior. That makes your use of what you thought of as praise to be a punishment or stimulus that stops a behavior. That is, it doesn't matter if you said good boy in a light-hearted and squeaky manner, to the dog, it may have sounded aversive. Which qualifies your use as an aversive, not a reinforcer. OC is still working as it should. It would be the person who assumed that what they did was a reinforcer and not an aversive. For example, when I say "down" to Shadow, I descend in pitch through the word and put a slight growl on it. To other humans, that might sound aggressive or like an aversive. It is not. It is a sound and inflection reminiscent of mother dog, who lowers her pitch at the end of sound to emphasize. It is a range that he can hear in the presence of other sounds. But it is not an aversive but a cue. A cue that is rewarded by rewards. So, I did not create a sustainable behavior with aversive. In that same vein, perhaps, at times, your "reinforcers" sound or seem like an aversive to the dog. What is reinforcer and what is aversive is determined by the dog's response to the stimulus or stimuli.

    As to the dopamine study, it is noted that there is long-term dopamine presence, as well. Which means it can be present during non-rewarding times, as well, though maybe not as much as during an event.

    Might it be likened to this. A creature suffers an injury and dopamine is released to ease pain. Later, dopamine, in the absence of pain also creates an euphoria, seemingly rewarding. In which case, dopamine has dual functions and isn't necessarily a sign of failure to learn through reward, conscious or unconscious.

    So, I notice in your response in Dog Star Daily to Dodman in re the fMRI that you suggest he has made too much out of the MRI and hippocampus study on memory and training yet you reject suggestions that you may make too much out of the dopamine study. "Do as I say, not as I do"?

    And maybe more germaine, from that response blog, your estimation of dog and wolf evolution sounds a little closer to mine, in that, at some point in the canid past, wolves and dogs descended different evolutionary paths. I still think that dogs did not descend from wolves. but came from similar ancestors, though there may have been a general protocanid out of which the different canids evolved.  For one thing, the skeletal structure of a dog, specifically the coronoid process on the zygomatic arch, is marked different in dogs than it is in wolves.Truly different species, in spite of similarities in one singular locus of mtDNA. It should also be noted that Wayne's mtDNA study was only with certain purebreeds and may not have accounted for any wolf-hybrids, intentional or accidental in the breeding. For example, if he had studied alaskan huskies, he would have found stronger convergence. I think it is the Huslia line of Alaskan Husky that includes, wolf, Siberian Husky, German Short-Hair Pointer, and Irish Wolfhound.

    • Gold Top Dog

    Hi Ron,

    Thanks for your response. You make some good points.

    ron2
    The nature of praise is irrelevant. What is important is how the dog responds to it. And reward and punishment are, as noted by many, maybe even yourself, operational terms, and should be free of emotional content for the human. So, you use praise and it stops a behavior. That makes your use of what you thought of as praise to be a punishment or stimulus that stops a behavior. That is, it doesn't matter if you said good boy in a light-hearted and squeaky manner, to the dog, it may have sounded aversive.

    It's possible, I suppose. It's difficult for an outsider (meaning someone who wasn't there at the time) to know. But I can assure that it was the most glorious kind of praise possible. In fact, I think you were closer to the answer when you were saying that the praise was more rewarding to Freddie than the piece of chicken was.

    This was a kind of watershed incident for me. First of all it made me wonder if the reason I'd found that some operant conditioning techniques I'd been testing didn't seem to work as well as "advertised" was that there might be some sort of underlying flaw in the overall theory and how it was constructed. Secondly, I had to ask myself, why did my praise stop the behavior that first time, and eventually extinguish it altogether?

    That's when I remembered a Harvard paper that was in the news at the time, showing that the human brain isn't designed to sustain feelings of happiness for long periods of time. I think they called it the "new car" or "Christmas morning" syndrome, or something like that. (In those days I was bad about keeping notes on these things.) The Harvard researchers said something to the effect that our brain chemistry tends to create a vague feeling of dissatisfaction that keeps us motivated to "get out of bed in the morning," and go out and do stuff.

    So I wondered if a dog's brain had a similar kind of chemistry. Remember, Freddie had eaten a full breakfast. So he wasn't scavenging because he was hungry for food. What was he hungry for? If a squirrel had wandered by, would he have dropped the chicken breast to stalk the squirrel? (I once popped a piece of chicken into his open mouth while he was staring intently at a squirrel, and he didn't take his eyes off the animal; he just casually spit the food right out of his mouth.)

    So that's when I came up with this idea that dogs have a drive to connect to things in the environment, which was inspired by my reading (in both sense of the word, as in "having read" and "interpreting";) Kevin Behan's book, Natural Dog Training.

    Since my praise made Freddie feel more connected to me than to the chicken breast, he dropped it and connected to me.

    To me, that explanation makes more sense than the idea that my praise was an aversive stimulus.

    ron2
    I notice in your response in Dog Star Daily to Dodman in re the fMRI that you suggest he has made too much out of the MRI and hippocampus study on memory and training yet you reject suggestions that you may make too much out of the dopamine study. "Do as I say, not as I do"?

     

    You may be right. However, fMRI scans are open to a lot of misinterpretation. The science is getting better, but there's also still a lot of free interpretation of the data. Also, and this takes us back full circle to the dopamine studies (not "study," singular, by the way), shortly before I'd written that reply to Dodman's search for the OCD gene in Doberman pinschers, I had come across a podcast interview with the author of a new book on learning and memory, Memory and the Computational Brain. His name was Randy Gallistel.

    In the podcast, Gallistel provides some compelling arguments that memory storage happens on a molecular, not neuronal or synaptic level. He says that "if you implement memory function at the molecular level, then you can put gigabytes worth of information into single cells; whereas, if you implement it at the synaptic level, you’re going to talking about kilobytes, at most." I found that interesting, and relevant, esp. in regards to the single, solitary fMRI study done on the hippocampi of London cabbies, which was part of Dodman's argument for the hippocampus being the sole (or primary) locus of memory storage in the brain.

    From what I can determine -- with my somewhat-limited capacity to understand such things -- the dopamine studies you're objecting to measure the production of that neurotransmitter in a much more exact manner (at a neuronal level), and do so in relation to the exact timing of rewards, and expected rewards, etc. So there are a number of factors that, I think, make my comment to Dodman not quite as blatant or black-and-white a matter of "do as I say, not as I do."

    Anyway, that's how I see it,

    LCK

    • Gold Top Dog

    You say that the MRI scans are open to a lot of interpretation. I hope that human patients don't hear that. They need to trust that their doctors are interpreting correctly in a diagnosis. Anyway, I come from a similar viewpoint in offering the interpretations I see of the dopamine study. They are also open to interpretations. And memory storage may happen on a molecular level, in a way similar to flip-flop circuits in computers ( a flipping state of amplifier stage turned either in full saturation or non-conductance, depending on the input signal, hence the use of binary as the root language of computers.) Since brain function has an electrical component to it, this might be how memory is stored, at least for a time. And, of course, we have own removable hard drives, namely written language.

    As I see it, the explanation for your dog's behavior is one of where OC is still working, just not in the pattern you expected. Either the praise was an aversive, or it was a cue, and that returning focus to you is more rewarding than staying locked on to whatever, which was rewarding, in the presence of drippy chicken. And I can see how, on the face of it, it would seem that the desire to connect to the environment, or something in the environment, such as a squirrel seemed stronger than the desire for immediate reward of offered chicken, an easier prey to get than a live squirrel.

    But the effects of reward and punishment do not require conscious thought. Merely that one is preferred over the other, as well as the dog's interpretation or reaction to the stimulus, and that can depend on neurological response. A member here had a Dogo de Argentino who could run through brambles and get pointy barbs stuck in his sac and come sauntering back to the house, these barbs still hanging there, not a care in the world. Which means he either wasn't registering it as pain or as pain connected with going through that particular patch. He wasn't connecting it as a punishment and as a punishment in connection with travelling that particular path. So, he would do it nearly every day. Because going through there meant something to him, in whatever drive he was in. On the surface, one could say that punishment doesn't work or punishment was working in reverse, reinforcing the behavior. That's not the case. What is the case is that the stimulus was not interpreted by the dog, consciously or unconsciously, and through whatever praticular neurology, as an aversive and/or as an aversive connected with the behavior or behavior chain. OC was still working but, as always, defined by the creature's response to stimuli.

    A reinforcer is a stimulus that tends to cause a behavior to be repeated. An aversive is a stimulus that tends to cause a behavior to not be repeated. A lack of reinforcer or weak reinforcer may cause a behavior to extinguish. A lack of aversive may cause a behavior to reinforce or repeat. In higher brain response, which is seen in dogs, one can withhold a reward to stop a behavior (-P). A behavior that reinforces by avoiding an unwanted stimulus is -R. A behavior that decreases or stops in the presence of an aversive is +P. A behavior that repeats or reinforces in the presence of a stimulus is +R. Any individual can respond to the same stimuli in different ways but the effect is the same. For some people, Deep Purple is way too loud and like to drive them away. For others, the sound is lovely and they will spend good money to hear something that is way too loud. Same stimulus, different response, all OC.

    • Gold Top Dog

    LCK - I finally finished reading the podcast interview with Gallistel. The compelling interview you speak of speaks more of the desire of Gallistel to switch learning theory to that of computational theory, ala the flip-flop circuits that form memory in a computer. And essentially processor that links to the binary states of chain molecules or neurons to abstract shapes and concepts such as smells, sounds, sights. Fair enough and I hinted as much with the reference to "Do Robots Dream of Electric Sheep?" and the consequent movie, "Blade Runner", which would ask the questions "Does a robot have a soul? Can he/she/it fear death? Love? Fight for one's own survival? At what point in the everincreasing complexity of connections and cellular numeration does sentience or self-awareness arise, and to what degree?" Is there a "tipping point" of one day being a protoplasmic mass of goo and the next generation, a tree shrew that remembers where it stashed acorns?

    In the interview, without explanation, he wishes to change the thinking of the scientific community away from "pavlovian response" or words to that effect in favor of computational theory of learning or brain activity. Why? I don't know. I don't see it as contradictory. In fact, I would think, he would see the learned response (associative) and the repeated learned response (operant) as supporting the information carried forward in time thing. He speaks of the bird experiment where birds, and others, can place a cache of resources, being there one time, and come back to exactly that same spot later in time to retrieve the items. That is a complex behavior chain but is a reward-based mechanism. How so? He who remembers where the cache is gets to eat and survive. That is, the ability to learn from the environment and have memory separates the living from the dead. And I don't see where that propose is opposed by whether our memory is stored as synaptic change or on a molecular level or even a sub-atomic level of states of electrons in the orbits around atoms, truly a binary or trinary function, like a computer.

    That is, Gallistel might be eventually right that memory is stored in an animal somewhat like a computer but that doesn't invalidate OC and in fact, better supports OC and possibly vice versa. Just as the MRI study on the hippocampus might very well be the smoking gun to physical changes that can be quantified as memory storage. To disallow Dodman his point might invalidate Gallistel's. Then, who are going to quote?

    Gallistel admits that his book is not a textbook nor approved as such but is, in so much of his words, an attempt to get future students to change their paradigm. Their initial viewpoint, which is not based on science or theories on evidence collected but is instead based on a pre-established theory and hopes that future students can then find the data to support the theory. That is not science, that is religion.

    What Newton did was science. What Einstein did was, at some points, religion. What Skinner did was science. It pre-supposed nothing about soul, abstract thought, merely what is the response of subject x to the environment. Out of that comes the learning theory of operant conditioning.

    As for the "probability" function of information intake, that would be due, I believe, to evolution. The animal that tends to sense that liquid stuff as water and has the tendency to drink it lives longer than the creature that doesn't. Nature is harsh but fair. You live or you die.

    Dopamine might even turn out to be a chemical that is simply an aid to neural behavior, which would explain why it is present in both reward and non-reward instances. That is, because it was present in earlier reward research, it was assumed at the time to be part of the reward sequence. Instead, it might merely serve as a marker than some event, rewarding or not, is being recorded. And does not invalidate OC.

    OC is simply a descriptive language that describes a physical event and outcome. Are their some who attach personal and religious meaning to it? Sure. That is the idiosyncracy of the individual, not a limitation or "failing" of the language of OC.

    Gallistel wants to put a stop to pavlovian thought. That is, in fact, throwing out the baby, the bath water, the whole durned bathtub. For an as yet unproven theory that answers some questions quite well and may even be anti-climactic, but throws away the newtonian thing that works just fine on the larger scale of what we do. In metaphor, it doesn't matter of the actual nature of the universe is sub-particles to even the quantum waveforms (that's the newest theories, as of 2010), it's still a newtonian world of apples falling on our heads.

    • Gold Top Dog

     Hi

    I generally take a back seat in these debates from LCK, pretty much

    1) So i don't get misquoted

    2) So that i don't spend hours trying to explain relatively complex facts on the back of a postage stamp which relatively speaking this forum is

    3) So i don't land up doing other people's research for them.

    The sad bit is that i would love to communicate with others on similar topics but without being rail roaded and diverted. I firmly belive that the possibility of convergence across many science disciplines (and non science) is possible, but needs to be done in a thougtful considered respectful way. I guess that i suffer with the paralysis of knowledge. It is pretty hard to be black and white and come across slick in a forum when you are used to being exact and careful to consider many factual rather than fictional views.

    So i will try to deal with facts alone. In the electronics world there is a myriad number of types of memory much of it classified as to whether it is able to be read from, to be written to, and how permanent and accesible it is to be. The range of structures is enormous. THere is AND and NAND memory. A new reintroduction is the use of ferro electric memory using nano particles.  It would not be a wonder if this memory in brain structures  was distributed as it can be in electronics. So the presence of memory or memory like structures in one area doesn't preclude the presence of other types of memory structure elsewhere. So there may be neuron type memory in one part of the brain and cellular type memory in another . Neuron type memory if one was to argue is cellular. http://en.wikipedia.org/wiki/Neuron "An elecricablly exicted cell".

    One type of pseudo memory are FAPS, or Fixed Action behaviours. We know that these circuits exist in the base of the brain. They may well be quite sparse, but they are quick. A similar feature exists in electronics. You don't often get dense and fast together. A bit like getting a ferrari that is the size of a bus, it ain't going to happen.

    The action of Dopamine is not fully known. In general and simple terms, Neuro Cheimcals act to excite Neurons, inhibit Neurons or do nothing with various sitmuli. So the presence of dopamine means zilch. What really matters is what else is there as well.

    Again, the wikipeida is not too far off the mark. I hesitate quoting this, i really don't want to deal with the misquotes and the tenuous use of it to build so called facts from very weak work

    http://en.wikipedia.org/wiki/Dopamine

    Again real time measurement of Dopamine is not really measurable, and the measurement is realtive and is intrusive. it interfers like hell with any behavioural experiment. http://www.neuroassist.com/dopamine-test.htm (BTW this is commercially biased but gives you an idea)

    The last but not least is that even if this was all true, it would not prove in itself that the process of OC and CC is wrong or non existent. it is a seperate arguement.

    I have often talked about abstraction and i guess that it is not understood. in a way i am talking in a much more genaralised way about he principles that underly OC. Here is a reference to a reasonable view of abstraction as i mean it 

    http://en.wikipedia.org/wiki/Abstraction_(computer_science)

    Lastly, OC is it's own proof. I am quoting Skinner here (again) "A science of behaviour need not wait until Neurology has done so (sic explained)"

    I guess the man was a hell of a lot more open minded than his detractors.

     Oh and while i respect and use Bheavourist techniques i could never be put in that corner. But i respect science and truth rather than smoke and mirrors based on "factlets" squeezed from really weird and often tenuous sources.

     I hope what i said is useful I don't profess to be a hugely knowledgeable practioner in the biological sciences, but i do work with neural networks on a day to day basis.

     

     

     

     

    • Gold Top Dog

    poodleOwned

     I generally take a back seat in these debates

    1) So i don't get misquoted

    2) So that i don't spend hours trying to explain relatively complex facts on the back of a postage stamp which relatively speaking this forum is

    3) So i don't land up doing other people's research for them.

     

     

    Well put.  Sharing the latest discoveries from practitioners and science research is commendable and enlightening activities of the forum.  But what must be respected in such discourse is that you have the right to state an opinion as nothing more than that, and not to insult participants' intelligence by arguing it as fact.  

    • Gold Top Dog
    Hi Ron,

    Thanks for your detailed and well-thought out response. I appreciate your taking the time to read the podcast transcription.

    ron2
    I finally finished reading the podcast interview with Gallistel. The compelling interview you speak of speaks more of the desire of Gallistel to switch learning theory to that of computational theory, ala the flip-flop circuits that form memory in a computer. And essentially processor that links to the binary states of chain molecules or neurons to abstract shapes and concepts such as smells, sounds, sights. Fair enough and I hinted as much with the reference to "Do Robots Dream of Electric Sheep?" ... In the interview, without explanation, he wishes to change the thinking of the scientific community away from "pavlovian response" or words to that effect in favor of computational theory of learning or brain activity. Why? I don't know.

    For those who haven't read the interview, and would like to be brought up to speed, here are the relevant passages:

    ---------------

    Dr. Campbell: Randy, could you go over what some of the key assumptions are that must be challenged, based on these experiments?

    Dr. Gallistel: Well, perhaps the most important thing is that when people are thinking about the nature of memory, they should stop thinking about the kind of Pavlovian conditioning experiments which have dominated neurobiological thinking about memory, and think about the demands on memory that are made by dead reckoning, and by this food caching behavior, and by innumerable other behavioral examples of the same kind.

    None of these examples is mysterious if you grant the animal read/write memory. All of these examples are not computationally mysterious. It wouldn't be difficult to simulate these performances in a conventional computer. But it's very important to understand that the reason it would not be difficult is that we understand very well how to represent locations, we understand very well how to encode food types, and so on.

    Most importantly, we understand how to encode that information physically in a computationally accessible memory in a conventional computer. There's no mystery about it. But if you try to do the same thing in a neural net, if you throw away the baby in the bath-namely, the read/write memory mechanism that is so fundamentally important in a modern computer-then trying to build a machine that can do what the jays are doing, or do what the ant is doing when it dead reckons, then it becomes extremely difficult.

    So, the most important thing is to focus our experimental and theoretical efforts on trying to imagine what the physical realization of such a mechanism-that is, a read/write memory-must be in nervous systems. We argue in the book that this mechanism is as basic to computation as genes are to life. Understanding this memory mechanism-the read/write memory mechanism-until we understand that mechanism, we have no hope of understanding how the brain computes, because that kind of memory is central to computation.

    This is the central argument of the book. And at the moment we do not know what that mechanism is in the brain. We argue that there's good reason to think that it will prove to be universal, just as DNA is universal. That is, that it performs a basic, simple, foundational function-namely, carrying information forward in time. There's no reason why a mechanism that works in one domain, or one context, wouldn't work in any other just as well. And therefore, there's no reason to assume that it's not a universal mechanism.

    Dr. Campbell: But you think we need to look beyond the synapse.

    Dr. Gallistel: Well, the question arises, so people are persuaded that it must be changes in synaptic conductance: could this mechanism be mediated by changes in synaptic conductance? Well, yes, it could be. But, for reasons we go into at great lengths in the book, it doesn't seem to us very likely that it is. It seems to us that once you grasp what the essential nature of the mechanism must be, then changes in synaptic conductance don't look very plausible or very attractive.

    It's not impossible. A change in synaptic conductance can be viewed as a change in the setting of a switch, or potentiometer. And that's what you need to implement any memory system. You need something that has more than one physically stable state. But we point out that synapses are way more complex than is actually required. I mean individual molecules have exactly that property.

    Like the rhodopsin molecule in vision. It's a perfect molecular switch. It has two different settings-an off setting, which is the normal setting, and an on setting, which is produced when a photon is absorbed by the molecule. Then that literally changes the physical configuration of the molecule-it isomerizes the molecule, in more technical language-and the isomerized molecule is now enzymatically active. That is, it's readable; it affects other physical processes within the cell.

    Another thing we try to get people to grasp is the difference in size between cellular-level structures, like synapses, and molecular-level structures, like rhodopsin. Because, if you implement memory function at the molecular level, then you can put gigabytes worth of information into single cells; whereas, if you implement it at the synaptic level, you're going to talking about kilobytes, at most. So, we argue that there are a variety of reasons-this isn't really the most important-but there are a variety of considerations that make changes in synaptic conductance not a very appealing story, once you realize that what we're looking for here is a read/write memory mechanism.

    ----------

    ron2
    I don't see it as contradictory. In fact, I would think, he would see the learned response (associative) and the repeated learned response (operant) as supporting the information carried forward in time thing. He speaks of the bird experiment where birds, and others, can place a cache of resources, being there one time, and come back to exactly that same spot later in time to retrieve the items. That is a complex behavior chain but is a reward-based mechanism. How so? He who remembers where the cache is gets to eat and survive. That is, the ability to learn from the environment and have memory separates the living from the dead.

    Okay, so first of all, I don't think Gallistel and others are trying to get rid of Pavlov's work. There's no question that this type of conditioning exists. The question is whether temporal contiguity is a necessary part of explaining how and why it works. And that's a pretty big question. (See: "Temporal maps and informativeness in associative learning," Balsam and Gallistel,Trends in Neurosciences, Volume 32, Issue 2, 73-78, 12 January 2009.)

    There's also no question that having the ability to remember things (like where you cached your food supply) is important to an animal's survival. The question is, how does this kind of behavior happen, what are the cognitive processing mechanisms?

    You seem to have a pretty good working understanding of how operant and classical conditioning work. And from what you tell us, you're using behavioral science techniques in training Shadow. So how many times would you say you've seen conditioning work to create new behaviors without the necessity for any repetition? And how can a behavior be learned via a reward-based mechanism, if said behavior is learned before it's even been rewarded? Remember, the blue jays come back to their cache before they get the reward of finding and eating it. Plus, I would say that this behavior seems pretty simple, and not at all part of a complex behavior chain, as you stated.

    So I think this brings us back into the territory of pattern recognition as opposed to after-the-fact, reward-based learning. I don't know of any types of conditioning that operate independently of the idea that behaviors are reinforced by positive feedback (reward) for their consequences. Yet you yourself said that this was an example of a "the information carried forward in time thing." How, in your view, would this feedforward type of learning gibe with operant conditioning?

    ron2
    Gallistel admits that his book is ... an attempt to get future students to change their paradigm. Their [King and Gallistel's] initial viewpoint, which is not based on science or theories on evidence collected but is instead based on a pre-established theory and hopes that future students can then find the data to support the theory.

    I didn't see any evidence of that at all. Where did you get that idea?

    ron2
    Dopamine might even turn out to be a chemical that is simply an aid to neural behavior, which would explain why it is present in both reward and non-reward instances. That is, because it was present in earlier reward research, it was assumed at the time to be part of the reward sequence.

    Current research suggests that dopamine plays a key role in enabling and motivating us to pay close attention to, and clearly remembering, certain key events, positive or negative.

    ron2
    And does not invalidate OC.

    No, but it calls one of its basic principles into question, i.e., the idea that there needs to be a strong temporal relationship between a behavior and its subsequent consequences in order for such a behavior to become learned.

    ron2
    Gallistel wants to put a stop to pavlovian thought. That is, in fact, throwing out the baby, the bath water, the whole durned bathtub.

    Again, I think that's a mischaracterization. What Gallistel actually said was, that when it comes to understanding how memory works, people need to stop thinking about the idea that there has to be a temporal relationship between an US and a CR. That's all. (And he's not suggesting throwing the baby out with the bath water; he's just pointing out that there are some holes in learning theory that aren't being addressed.)

    Thanks again,

    LCK

     

     

    • Gold Top Dog

    Lee Charles Kelley
    No, but it calls one of its basic principles (that there needs to be a temporal relationship between a behavior and its subsequent consequences in order for such a behavior to become learned) into serious question.

     

    Actually, to characterise the temporal relationship between behaviour and consequence as one of its basic principles isn't completely accurate. To be more precise, it is contingency, that is, the dependency of the conseequence following the behaviour (i.e. the association/relationship) that is important, rather than the actual temporal sequence of behaviour.

     This has been evidenced in backwards conditioning experiments (where the conditioned stimulus follows the unconditioned stimulus), and trace conditioning experiments (where there is a delay between the conditioned stimulus and the unconditioned stimulus), which still demonstrate conditioning/learning.

     

     

     

     

    • Gold Top Dog

    poodleOwned
    I generally take a back seat in these debates from LCK, pretty much

    1) So i don't get misquoted

    2) So that i don't spend hours trying to explain relatively complex facts on the back of a postage stamp which relatively speaking this forum is

    3) So i don't land up doing other people's research for them.

    I agree that it can be very difficult to stay actively engaged, not to mention positive-minded, for those very reasons.

    poodleOwned
    The sad bit is that i would love to communicate with others on similar topics but without being rail roaded and diverted. I firmly belive that the possibility of convergence across many science disciplines (and non science) is possible, but needs to be done in a thougtful considered respectful way. I guess that i suffer with the paralysis of knowledge. It is pretty hard to be black and white and come across slick in a forum when you are used to being exact and careful to consider many factual rather than fictional views.

    I hope that wasn't directed at me. I do my best to be thoughtful and respectful. And I don't believe I've ever railroaded you or anyone else here. And my views are far from fictional; they're based on careful research. I rarely post anything that doesn't have a sound scientific basis.

    And I don't think I've ever come close to railroading anyone about anything. If I have, give me some examples.

    poodleOwned
    So i will try to deal with facts alone. In the electronics world there is a myriad number of types of memory much of it classified as to whether it is able to be read from, to be written to, and how permanent and accesible it is to be. The range of structures is enormous. THere is AND and NAND memory. A new reintroduction is the use of ferro electric memory using nano particles.  It would not be a wonder if this memory in brain structures  was distributed as it can be in electronics. So the presence of memory or memory like structures in one area doesn't preclude the presence of other types of memory structure elsewhere.

    I agree 100%, which is why I took exception to Nicholas Dodman's view on the hippocampus.

    poodleOwned
    One type of pseudo memory are FAPS, or Fixed Action behaviours. We know that these circuits exist in the base of the brain. They may well be quite sparse, but they are quick.

    Are you discussing FAPS as a form of genetic memory? Because if not, to classify them as a form of memory doesn't make sense. At least not to me. 

    poodleOwned
    The action of Dopamine is not fully known. ... Again real time measurement of Dopamine is not really measurable, and the measurement is realtive [relative] and is intrusive. it interfers like hell with any behavioural experiment.

    I don't think I ever said that dopamine is fully understood. As for real-time measurements, the one study I cited a reference to, involved measuring activity in specific dopamine neurons, in real time, not the amounts of the amounts of the substance found in the test subject's urine later on.

    poodleOwned
    i really don't want to deal with the misquotes and the tenuous use of it to build so called facts from very weak work

    Examples?

    poodleOwned
    Lastly, OC is it's own proof. I am quoting Skinner here (again) "A science of behaviour need not wait until Neurology has done so (sic explained)" I guess the man was a hell of a lot more open minded than his detractors.

    There are two ways to look at Skinner's statement. Yours, which seems to be in favor of the idea that behavioral scientists don't need to  waste time studying neurobiology, and the opposite, which is that it's not a waste of time at all. Temple Grandin's version of that quote (something he reportedly said to her personally), was "We don't need to understand how the brain works. We have conditioning."That hardly puts him in a good light. And I have to say, that in all my research into Skinner, I have never seen any evidence that he was close to what you could call, "open minded." In fact, I think most scientists would tell you that he was just the opposite, at least where conditioning was concerned.

    poodleOwned
    Oh and while i respect and use Bheavourist techniques i could never be put in that corner. But i respect science and truth rather than smoke and mirrors based on "factlets" squeezed from really weird and often tenuous sources.

    I empathize with your frustration. It's not easy to wade through all the material in a forum like this, coming in from various sources, and being able to keep one's focus on the salient parts of other people's arguments. But I'd like to point out that using phrases like "smoke and mirrors," "faclets," and "tenous sources," provides no real information to the discussion. 

    So while I know that this is going to sound enormously self-serving, I can either slough off your comment about smoke and mirrors -- which isn't true -- or I can defend myself. So in this case I'm going to risk sounding self-serving and defend myself.

    I've gotten positive feedback on my PsychologyToday.com blog articles from half-a-dozen PhDs, one of whom is currently doing work on canine cognition, and has expressed an interest in working with me on co-writing a study or two. Another asked me for help on an article he's writing about whether guide dogs know their owners are blind. Two evolutionary biologists gave me some very complementary comments about my hypothesis that oxytocin may have been a key factor in the development of the pack instinct in wolves. A very well-known animal behavior expert called me up and suggested we have coffee the next time he's in New York. Still another -- a woman I've known for 25 years -- only recently came across some of my articles and said she's amazed and impressed at how rigorous my research is.

    Now, I'm not saying any of this to prove anything to you or anyone else. I'm frankly amazed that these people have said such kind things to me. And I can already hear the snarky comments to the contrary. But my purpose here is not self-aggrandizement, it's discussing the issues. So if you don't mind, I'd ask you please, if you have specific problems with my data, please be so kind as to let me know that those problems are, specifically, so that I can address them.

    poodleOwned
      i do work with neural networks on a day to day basis.

    That's interesting. In what capacity?

    LCK
    • Gold Top Dog

    Lee Charles Kelley
    I rarely post anything that doesn't have a sound scientific basis.

     

     

    Actually LCK , on subjects that i know a hell of a lot about you often get the sense really wrong. So wrong in fact that i let it slide. It can be very difficult to start with what you say and get it back to something that is closer to the truth. I definitely don't think it is deliberate, i just guess that the theory base in my area is not there.

    Lee Charles Kelley
    Are you discussing FAPS as a form of genetic memory? Because if not, to classify them as a form of memory doesn't make sense. At least not to me. 

     

     Actually read only memory and a state machine can give equivalent results. In our little world we go with what costs less and which does the job. They are quite different. A state machine responds to stimuli and uses a varity of prewired gates. Commonly we use AOI (AND OR INVERT) structures or the inverse. Memory is a subtly different structure and often needs energy. It can be wired so that it mimics a state machine. In this case a stimulus adressess a space in memory which contains the response.

    Lee Charles Kelley
    in real time,

     

    The nearest we have to real time measurements would involve intrusive sampling via either blood samples or salivia samples. Then you are down the rocky road that behavourism visits.

     

    Lee Charles Kelley
    i do work with neural networks on a day to day basis.



    I don't wish to use this forum to promote commercial products but we use neural networks to reduce the complexity of the silicon involved in some energy saving products. Neural Networks do not preclude the loss of memory, anything but. This article here is a good start http://www.learnartificialneuralnetworks.com/ .Unfortunately you do need the maths.

    Often we restrict our thinking to Synchronous single operation machines when we make comparisons. There are several other forms. Paralell machines are now much more common. I can buy a machine that does paralell operations for around $2 . My guess is that you PC does some paralell operations too.

    We have found that using  Asynchronous neural networks have some huge complexity advantages. If i said that they were easy to understand and get going i would be lying they aren't. We just don't have the tradition of doing this. It goes against everything that we were ever taught.

    They are small and quick and energy saving. Please don't stretch this!

    Now i am not an expert in a lot of what else you say but sometimes i am a little open mouthed at what you do say. I remember the example of the dog and chicken. I thought it was common practice to do what you found exceptional, at least it is here.

    Your Hypothesis  that oxytoxin is the reason for pack drive in Wolves seems self evident and quite old  to me. So it may not be your hypothesis. Again, it is a hell of a lot fo work to back reference this.

    Another is the idea that back conditioning may work . The relevant timing of the reward in conditioning has proverbially had the b**m studied out of it. There is a strong correaltion between appropiate timing and the effectiveness of the reward. There are weak statistical links that could be used if a very heavy barrow was being pushed up hill that suggest that other condtioning might be effective.

    I have stated that i am interested in the time sequence of many of these early experiments but believe that the averaging was important to get some useful results in the time that the experiments were performed. 

     

     

    • Gold Top Dog

    You insist there is scientific research backing your hunch-du-jour postulated as a factual tenet of your energy theory, BUT REFUSE TO SHOW SUCH CORROBORATING REFERENCES.  Not only that, but pressed by an undeniable charge of absurd theorizing on your part, you deflect for the moment, only later to come back saying you never held such a position (QM, anyone). In all due respect to Kevin Behan, we should probably quit referring to your ruminations  as NDT, since you now deny most of its major principles.

     

     

    Lee Charles Kelley

    I've gotten positive feedback on my PsychologyToday.com blog articles from half-a-dozen PhDs, one of whom is currently doing work on canine cognition, and has expressed an interest in working with me on co-writing a study or two. Another asked me for help on an article he's writing about whether guide dogs know their owners are blind. Two evolutionary biologists gave me some very complementary comments about my hypothesis that oxytocin may have been a key factor in the development of the pack instinct in wolves. A very well-known animal behavior expert called me up and suggested we have coffee the next time he's in New York. Still another -- a woman I've known for 25 years -- only recently came across some of my articles and said she's amazed and impressed at how rigorous my research is.

    I'm not saying any of this to prove anything to you or anyone else. I'm frankly amazed that these people have said such kind things to me.

     

     

    I, too, am amazed people said such nice things to you.  Scienticic fact is not equivalent to supposed compliments.  I would not accept such loosely worded hearsay as fact unless presented with each correspondence and the context in which it occurred.  On second thought, we have no reason to accept even this, given your history of editing your combox comments to put forward your own view, instead.

     Despite your delusion to the contrary, what you've implied above is highly improbable as stated- you have NO PROFESSIONAL CREDENTIALS to help a cognitive researcher - your name on an article submitted for peer review to a professional journal would kill it outright.

     

    We are not stupid, lee.