It depends on what you mean by context, as well as discrete cue. I think the obvious answer is yes, since darkness, for example, is a lack of visual context. Any discrete cue added to this null context will provide context. Adding more cues will build context. If a discrete cue adds to context it serves as context. Have a look at my stimulus control paper available on ResearchGate. It might help.

Hi Rodrigo,

I do not think that a discrete cue can be used as context, as context is by definition the sum of cues generated by the environment perceived by the animal. So I would tend to answer no to your question. Now the reinstatemant is basically the re-expression of a response to the CS after extinction in response to exposure to the US and the conditinoing context (in absence of the CS). In that definition, context is actually again not a discrete cue. You could imagine to have a CS-US association with an additional 2nd stimulus that is not associated with the US, and see if the 2nd CS can be used to reinstate the response to the 1st CS. I am not expert enough in conditioning paradigm to know if this has been done by someone. If this is blocking you i can ask to some expert i know.

best

Hi, Jonas. Thank you for your response. Maybe the exposure to the US in the paradigm that you describe could be seen as a "molar" (non-discrete) operation; nevertheless, each presentation of the US can qualify as a discrete operation. Certainly, what promotes the re-expression of the conditioned response is the sum of all the presentations of the US, but I think that each presentation counts. Now, I was aware of that paradigm, but I'm looking for a study in which, as your hypothetic example describes, a CS alone promotes the re-expresion (or maintenance) of a conditioned response. Maybe asking an expert would help.

Of  course, a discrete cue can be used as context. A discrete cue can be made to be the ENTIRE context under special circumstances. For example,  you could use a set of training spaces, arenas of different shapes that the subject has previously explored. Let's say that one is bounded on all sides by walls and another by a precipitous drop-off. Each is associated with, say, a continuous audio tone (CS), which differs in pitch between the spaces, thus providing context. You introduce a previously acclimated subject into one of these spaces, only this time in the dark or blindfolded and the subject, of course, would use the discrete cue as context to determine which space it was in (I'm quite certain that mice would do this). With an appropriate experimental design, one would be able to confirm this through the UR, which might be cautious movements in the drop-off condition. One might argue that, in this example, the floor and it's edges could also serve as context, but there are ways to control for that, such as only scoring behavior  in the center of the arena.

I played around with something similar with mice in my Barnes maze experiments (see my posted papers), but I used spatial cues that were not necessarily discrete. What I described above represents a more rigorous test of discrete cue context, but my paper might help you think about your own questions.

it might help to have some knowledge of your research subjects and your goals.

Hi, Randall. Thank you for your insights. I'm currently studying rat behavior in conditioning chambers. The main point is: I trained a speciphic behavior in a condition in which a discrete cue (lets say a light) is occassionally presented. Then, I tested the putative learning process formerly trained maintaining the occassional presentation of the discrete cue. I just wanted to empirically justify my (and perhaps yours as well) intuition that the presentation of the cue could help to maximize the meassurable amount of the learning process I'm assessing. I hope that makes sense.

I suppose so, somewhat anyway. I'm curious about your methods. Are you using operant chambers with lighted nose-poke holes? Are your subjects in a lit area or is it dark when the cue is presented. If it's dark then light may not represent a discrete cue, since it would also provide spatial context.

The operant chambers are dark through the session, except when the light is presented. I defines "discrete cue" as a cue that has a well defined start and end. From a temporal point of view, 8 seconds light presentations are certainly discrete. Conversely, from a spatial point of view, as you mentioned, it's not easy to conceive lights as a discrete cue. Maybe I was not clear in phrasing my question. Perhaps, the main point is if there is evidence of an elemental (monosensory) intermittent cue with sufficient power to produce effects qualitatively comparable to the ones in the context (e.g., Mark Bouton's group) literature.

Hi Rodrigo, Randall,

Interesting discussion, Randall, indeed the only way to get a discrete cue to become the context would be this clue to be the only perceptible signal by the rodent. However that would mean finding a way to ensure absolutely no other sensory stimulus is perceived by the animal, which I think is almost impossible. Even if you play with the ground to be randomly regularly changed, the mice will still integrate all sensory input to generate a context, except if you sensory deprive them (blinding, tactile, olfactory deprivation would be needed if you want only a tone defined context). 

One could attempt to render all other modality meaningless by presenting them in random combination with the CS, however that will not remove them from the context but simple render them as a confounding factor in the encoding/recall in my opinion.

Rodrigo, please let me know if you proceed with experiments, i would be curious of the design you come up with as well as any results.

kind regards

Jonas

I have to respectfully disagree with you, Jonas. Context is not merely the sum of all perceptible signals. It is better characterized as the actionable cues in a given environment--the cues that can influence behavior by way of some meaningful assessment of the environment.

Let me unpack this thought, since it seems to be of interest to some folks. Say we have a set of arenas, all of different shapes and all with smooth floors. If we place a subject into one of these arenas in the dark, the floor provides sensory feedback, but no context about which arena is being occupied. It does, however, provide a discrete contextual cue, a supporting surface, about which nothing else is known without further action by the subject. This is a single-cue context, regardless of whether or not it can be perceived via multiple sensory modalities.

Of course, one might try to argue that darkness is part of the context. After all, it influences behavior (inducing caution). Can a lack of input contribute to context? No, it cannot. It is the lack of context that the subject fears, not the lack of light.

Now let us add a sound, say an audible hum with which the subject is familiar. Let the sound be continuous from the time the subject is introduced to the arena. Did the sound add anything to the context? No, because it did not provide any actionable information. It is just as though the sound were produced by tinnitus, which draws attention to a can-of-worms regarding internal cues (some of which provide context while others don't). I am purposefully using this scenario to contrast what I wrote in an earlier post about sound AS context. Context is about the organism (or rather the organization of an organism's brain) as much as the environment. There are numerous potential (perceivable) signals that occur in environments that provide no context. Indeed, we know that brains filter out much of the available information in order to construct a clear context--an actionable perception of the environment.

I hate to be too wordy or pedantic on this forum, so I hope you see where I'm going here. We have to acknowledge that context, while generated by the environment, is the product of a nervous system. As such, a single cue can become the entire context through processes like bias, salience, and filtering. I acknowledge that single cue context might be rare--I think it is--but far from unachievable in a controlled experimental setting.

Rodrigo: The lights are reasonably discrete cues in your setup, and they do not really provide any additional spatial information (these are rodents--they get good spatial info from their whiskers). The light itself probably does not provide any actionable information about the box that your subjects don't already possess, but if the light signals a reward or provides actionable information about an event, THAT is context.

Hi, Jonas; Hi, Randall.

I think that both conceptualizations of context are logically valid: (a) as all sensory input for an individual, and (b) as any cue that provides actionable information in a given environment. I thank you both for your time and effort in elaborating your answers. Maybe each definition would be particularly usefull in different situations. Certainly, every reader of this thread will gain a valuable insight.

By now, I'm not systematically testing this hypothesis of a 'discrete' cue functioning as context. Rather, I crafted my research protocol based on this assumption hoping it would maximize the amount of measurable effect of the administered treatment. The treatment (Pavlovian conditioned inhibition confirmed by summation test) worked nicely, but unfortunately I don't know the relative contribution of the putative contextual discrete cue, as I have no control groups.

Hi Rodrigo, Jonas, and Randall,

thank you for this [old] conversation!

I would accentuate the definition of context you provided, as a summation of sensory stimuli/input that have an equal probability to occur during "US-intervals" and "non-US intervals", while a discrete CS, by definition, is to appear at higher probability in association with one of two. Thus, I would argue with the conclusion that context can be defined " (b) as any cue that provides actionable information in a given environment", suggesting it is rather "a given environment, capable to assign specific actionable information to any cue", e.g. "occasion setting", as described by Crombag et. al., 2008.

Hi, Koral. Thank you for this addendum; it certainly tack on an important point to the conversation, which I think pretty much rounds it up! It is reasonable to consider that a contextual cue, in order to be conceived as such, must have an equal probability during the US and during the absence of the US. This is not the case, by the way, in my protocol, given that the discrete cue is alternated with the target trials. That makes the discrete cues to be negative related with the target trials; ergo, it have an uneven probability in the presence vs in the absence of the US.

Thank you again for your insight and for the recomended reading.

Kind regards

R.Sosa