Output details

Abrupt transitions between prefrontal neural ensemble states accompany behavioral transitions during rule learning.

<24> First study ever providing evidence at the single unit level that learning behavioral rules is not at all a gradual process, but is accompanied by abrupt phase transitions in neuronal dynamics. This implies that animal learning is not properly described by reinforcement learning, but rather by active decision making among candidate rules. This study led to large follow-up funding through DFG, a study in Science, widely triggered press releases and was featured in a Nature comment, due to its fundamental implications for animal learning. The work also has important implication in computational reinforcement learning.