Developing Intelligence

Though anatomically heterogenous, the human prefrontal cortex seems to perform a rather general function: it actively maintains context representations to guide and control behavior. What, then, is the reason for the anatomical diversity within this region of the brain?

Some theories suggest that prefrontal cortex (PFC) is organized to represent increasingly protracted contexts. According to this “temporal cascade” model, the most posterior regions of PFC are responsible for maintaining only the most current contextual information (e.g., “the doorbell rang”) whereas progressively more anterior regions of PFC are responsible for more temporally delayed context information (e.g., “you’re at your friend’s house”).

What evidence supports this view? As reviewed by Koechlin & Summerfield’s in-press TICS article:

- In a task-switching paradigm, more anterior regions of PFC show increased blood oxygenation to stimuli with a long temporal impact (instructional cues specifying which task to perform) whereas more posterior regions of PFC show selective sensitivity to stimuli with a more immediate temporal impact (for example, the number of possible responses within a given block of trials). The informativeness of each type of stimulus (quantified by Shannon information) predicts reaction time in a way that is statistically independent from each other type of stimulus – confirming that the “temporal cascade” theory provides a good mapping between cognition and neuroanatomy in the PFC. See this post for a more detailed description of this study.

- The most anterior region of PFC – BA 10, a region I’ve called the “anterior frontier” – seems to be specialized for even more complex operations over time, loosely referred to as “cognitive branching.” Cognitively speaking, this is a process by which multiple if-then rules might be structured at an even higher level. Indeed, Koechlin et al previously showed that BA10 is more active when one task must be temporarily suspended in favor of another one, relative to a similarly complex situation in which one task must be kept active across a delay.

- Information relevant across longer timescales (which is thus maintained in more anterior regions) may actually be “downloaded” to more posterior regions! Koechlin et al have shown that activity in these prefrontal regions is temporally correlated (known as “functional coupling”) only as predicted by this “temporal nesting” theory. That is, all PFC regions are more coupled in conditions where more temporally protracted information is being provided, whereas only more posterior regions are functionally coupled when more immediate information is being provided.

- The authors briefly review evidence that controlled retrieval processes recruit more anterior regions of ventrolateral PFC (vlPFC), whereas “postretrieval mechanisms” appear to be engaged by more posterior regions of vlPFC. This is consistent with a perspective in which anterior vlPFC represents contextual information, triggering “pattern completion” processes in the service of retrieval. More posterior regions of vlPFC may then represent or select more specific or “temporally precise” information to further hone this retrieval process.

- Broca’s area, long thought to be the “grammatical module” of the brain, appears to show increasing activation with the complexity of “structured behavioral plans,” according to evidence briefly reviewed by Koechlin et al (in which Broca’s area was defined as BAs 6, 44, and 45, the complexity of behavioral plans was manipulated as the difference between simple, overlearned S-R mappings and a task-switching paradigm.) This study was also reviewed reviewed by John Hawks, from a more evolutionary perspective.

While the temporal cascade model of prefrontal function is not universally accepted, these results do provide substantial support for the model. At the same time, this seems like a suspiciously elegant and parsimonious explanation of prefrontal organization; what about functions like reversal learning? “Inhibition”? Monitoring? Attentional reorienting? Conflict detection/resolution? Clearly, this theory has far to go before it can explain all the evidenced functions of prefrontal cortex, but it does seem like a strong contender.