A cognitive neuroscience perspective on insight as a memory process: encoding the solution

This is one of two chapters on "A cognitive neuroscience perspective on insight as a memory process" to be published in the "Routledge International Handbook of Creative Cognition" by L. J. Ball & F. Valleé-Tourangeau (Eds.). While the previous chapter discussed the role of long-term memory for solving problems by insight [https://psyarxiv.com/zv4dk], the current chapter focuses on the role of insight problem solving for long-term memory formation. Insight in problem solving has long been assumed to facilitate memory formation for the problem and its solution. Here, we discuss cognitive, affective, and neurocognitive candidate mechanisms that may underlie learning in insight problem solving. We conclude that insight appears to combine several beneficial effects that each on their own have been found to facilitate long-term memory formation: the generation effect, subjective importance of the discovery of the solution, intrinsic reward, schema congruence, and level-of-processing. A distributed set of brain regions is identified that is associated with these processes. On the one hand, the more affective response related to pleasure, surprise, and novelty detection is linked to amygdala, ventral striatum, and dopaminergic midbrain activity, supporting an important role of reward learning. On the other hand, insight as completing a schema is associated with prior knowledge dependent and medial prefrontal cortex mediated memory formation. Thus, learning by insight may reflect a fast route to cortical memory representations. However, many open questions remain, which we explicitly point out during this review.

Markers of Insight

Sometimes creative ideas come to mind following a step-by-step conscious reasoning, other times they rush into consciousness unexpectedly as sudden insights. Research on insight problem solving began about a century ago with a chimpanzee having an Aha! moment on how to pile up some boxes to reach a banana that was hanging from the ceiling (Köhler, 1917). Thanks to the development of neuroscientific techniques, researchers gained a better understanding of the physiology associated with insight, supplemented classic theories, and raised new questions about the cognitive processes involved in it. With the final goal of sketching a comprehensive understanding of the neurocognitive bases of insight, this chapter describes and updates the knowledge we gained about its functioning. A review of the last two decades of research on the ‘markers of insight’ revealed that: a) insights are paired with an internal attention allocation; b) the right anterior temporal lobe is a key node for insights, and if stimulated the frequency of insights increases; c) the feeling of pleasure and excitement that accompanies insights is warranted by the involvement of the reward-dopamine system; d) pupil dilation marks the switch into awareness of Aha! moments. Taken together these results indicate that insight below awareness processing might be explained by the involvement of subcortical areas responsible for learning, alertness, and emotions which are evolutionary more ancient than the cortex and it involves areas of the cortex responsible for information integration presumably together/after the switch into awareness. In conclusion, I summarize these points in terms of the defining characteristics of having an insight.

Cognition stays wild: A commentary on Ross and Vallée-Tourangeau’s Rewilding Cognition

A “failed” experiment (Ross & Vallée-Tourangeau, 2021) tried to reveal the role played by materiality in solving an insight problem that made reference to embodied action, leading to valuable insights about the nature of cognition and the experimental method. In this commentary, we argue that this study reveals various forms of interactivity and brings new evidence against the idea that “pure” cognition can be isolated from either materiality or sociality. The question becomes, then, not whether the use of objects helps or hinders problem solving, but how objects, bodies, and other people participate in it, even in controlled lab settings, and to what effect. Reflections are offered on why and how cognition stays wild (i.e., embodied, dialogical, and surprising) and what this means for experimental work.

Between automatic and control processes: How relationships between problem elements interact to facilitate or impede insight

Solving a problem requires relating the pieces of information available to each other and to the solution. We investigated how the strength of these relationships determines the likelihood of solving insight tasks based on remote associates. In these tasks, the solver is provided with several cues (e.g. drop, coat, summer) and has to find the solution that matches those cues (e.g., rain). We measured the semantic similarity between the cues and the solution (cue-solution similarity) as well as between cues (cue-cue similarity). We assume those relationships modulate two basic processes underlying insight problem solving. First, there is an automatic activation process whereby conceptual activation spreads across a semantic network from each cue node to their associated nodes, potentially reaching the node of the solution. Thus, in general, the higher cue-solution similarity, the more likely the solution will be found (Prediction1). Second, there is a controlled search process focused on an area in semantic space whose radius depends on competing cue-cue similarity. High cue-cue similarity will bias a search for the solution close to the provided cues because the associated nodes shared by both cues are highly co-activated. Therefore, high cue-cue similarity will have a beneficial effect when the cue-solution similarity is high but a detrimental effect when cue-solution similarity is low (Prediction2). Our two predictions were confirmed using both verbal and pictorial remote association tasks, supporting the view that insight is dependent on meaningful relationships between cues and solutions, and clarify the mechanisms of insight problem solving in remote associates.

Explanations in the wild

Why do some explanations strike people as highly satisfying while others, seemingly equally accurate, strike them as less appealing? We analyze thousands of open-ended explanations generated by lay-people in response to 'Why?' questions spanning multiple domains, to discover (1) what kinds of features are associated with ratings of explanation quality; (2) whether people can tell how good their explanations are; and (3) which cognitive traits predict the ability to generate good explanations. Our results support a pluralistic view of explanation, where satisfaction is best predicted by either functional or mechanistic content. Respondents were better able to judge how accurate their explanations were than how satisfying they were. Insight problem solving ability was the cognitive ability most strongly associated with the generation of satisfying explanations.

Aha! under pressure: Is the Aha! experience constrained by cognitive load?

The Aha! moment‒ the sudden insight sometimes reached when solving a vexing problem‒ entails a different problem-solving experience than solution retrieval reached by an analytical, multistep strategy (i.e., non-insight). To date, the (un)conscious nature of insight remains debated. We addressed this by studying insight under cognitive load. If insight and non-insight problem solving rely on conscious, effortful processes, they should both be influenced by a concurrent cognitive load. However, if unconscious processes characterize insight, cognitive load might not affect it at all. Using a dual-task paradigm, young, healthy adults (N = 106) solved 70 word puzzles under different cognitive loads. We confirmed that insight solutions were more often correct and received higher solution confidence. Importantly, as cognitive load increased, non-insight solutions became less frequent and required more solution time, whereas insightful ones remained mostly unaffected. This implies that insight problem solving did not compete for limited cognitive resources.

The Cognitive Underpinnings of Multiply-Constrained Problem Solving

Individuals encounter problems daily wherein varying numbers of constraints require delimitation of memory to target goal-satisfying information. Multiply-constrained problems, such as the compound remote associates, are commonly used to study this type of problem solving. Since their development, multiply-constrained problems have been theoretically and empirically related to creative thinking, analytical problem solving, insight problem solving, and a multitude of other cognitive abilities. In the present study, we empirically evaluated the range of cognitive abilities previously associated with multiply-constrained problem solving to assess common versus unique predictive variance (i.e., working memory, attention control, episodic and semantic memory, and fluid and crystallized intelligence). Additionally, we sought to determine whether problem-solving ability and self-reported strategy adoption (analytical or insightful) were task specific or task general through the use of novel multiply-constrained problem-solving tasks (TriBond and Location Bond). Performance across these tasks was shown to be domain general, solutions derived through insightful strategies were more often correct than those derived through analytical strategies, and crystallized intelligence was the sole cognitive ability that provided unique predictive value after accounting for all other abilities.