Context-dependent learning and causal structure

Abstract

The notion of "context" has played an important but complicated role in animal learning theory. Some studies have found that contextual stimuli (e.g., conditioning chamber) act much like punctate stimuli, entering into competition with other cues as would be predicted by standard associative learning theories. Other studies have found that contextual stimuli act more like "occasion setters," modulating the associative strength of punctate stimuli without themselves acquiring associative strength. Yet other studies have found that context is often largely ignored, resulting in transfer of performance across context changes. This article argues that these diverse functions of context arise in part from different causal interpretations of the environment. A Bayesian theory is presented that infers which causal interpretation best explains an animal's training history, and hence which function of context is appropriate. The theory coherently accounts for a number of disparate experimental results, and quantitatively predicts the results of a new experiment designed to directly test the theory.

Plans, Habits, and Theory of Mind

“…For example, context-dependence, like habitual action, is also sensitive to repetition [75]. Interestingly, recent work suggests that habits are more context-dependent than plans [76].…”

Private-public mappings in human prefrontal cortex

“…In these scenarios, context may be particularly important for adaptive behaviour. In psychology, context is defined as the set of circumstances around an event (e.g., we may hear a ringing phone in different environments such as our home or a friend's house) (Gershman, 2017;Maren, Phan, & Liberzon, 2013). Representing context is useful because it may carry information about what to expect or what to do (e.g., we should answer a ringing phone at home but not at a friend's house) (Gershman, 2017;Maren et al, 2013).…”

“…In psychology, context is defined as the set of circumstances around an event (e.g., we may hear a ringing phone in different environments such as our home or a friend's house) (Gershman, 2017;Maren, Phan, & Liberzon, 2013). Representing context is useful because it may carry information about what to expect or what to do (e.g., we should answer a ringing phone at home but not at a friend's house) (Gershman, 2017;Maren et al, 2013). Context may also hold structural information that can be generalised across events (e.g., many behaviours which are acceptable at home are not acceptable when visiting others) (Gershman, 2017).…”

“…Representing context is useful because it may carry information about what to expect or what to do (e.g., we should answer a ringing phone at home but not at a friend's house) (Gershman, 2017;Maren et al, 2013). Context may also hold structural information that can be generalised across events (e.g., many behaviours which are acceptable at home are not acceptable when visiting others) (Gershman, 2017). There is evidence that such structural knowledge is encoded in the medial temporal lobe and ventral PFC (Behrens et al, 2018;Niv, 2019).…”

“…First, how are context-dependent private-public mappings learned? Model-free reinforcement learning (e.g., Q-learning) (Sutton & Barto, 1998) may be able to learn the value of taking a public action given a private state and a social context but is likely to be too memory intensive (e.g., storing Q-values for every state-action pair in every context) and too slow (e.g., having to experience every state-action pair in every context) without model-based components (e.g., using a model of the world to generalise across state-action pairs or across contexts based on their similarity) (Gershman, 2017). Second, what happens when there is uncertainty about the current context?…”