Back

Too much or not enough? Optimal level of human intracranial theta activity for rule-switching in the subthalamo-medio-prefrontal circuit

Laquitaine, M.; polosan, m.; Kahane, p.; Chabardes, s.; yelnick, j.; Fernandez-Vidal, S.; domenech, p.; Bastin, J.

2023-12-12 neuroscience
10.1101/2023.12.11.571129 bioRxiv
Show abstract

The ability to strategically switch between rules associating stimuli and responses as a function of changing environmental demands critically depend on a neural circuit including the dorsomedial prefrontal cortex (dmPFC) and the basal ganglia. However, the precise neural implementations of rule switching remain unclear. To address this issue, we recorded local field potentials from two groups of rare patients performing a rule-switching paradigm: (1) deep brain recordings of the subthalamic nucleus (STN) in patients with obsessive-compulsive disorder, and (2) stereo-electroencephalogram from dmPFC of drug-resistant epileptic patients. We fitted a hierarchical drift-diffusion model (HDDM) to patients choice behavior and found that rule-switching was associated with a shift in the starting point of evidence accumulation (z), effectively disentangling rule switches from the selection of a new response. At the neural level, we found that theta band (5-10 Hz) activity increased in dmPFC and STN during switch compared to non-switch trials, while temporally delayed and excessive levels of theta activity led to premature switch errors. This seemingly opposing impact of increased theta rhythms in successful and unsuccessful switching could be explained mechanistically using a neural HDDM, as trial-by-trial fluctuations in theta power negatively correlated with the subjects starting point parameter. Together, these results shed a new light on the neural mechanisms underlying the rapid reconfiguration of stimulus-response associations, revealing a Goldilocks effect of theta band activity on rule switching behavior.

Matching journals

The top 4 journals account for 50% of the predicted probability mass.

50% of probability mass above

"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.