Neural representations supporting reuse of abstract structural knowledge in humans

A human ability to adapt to the dynamics of novel environments relies on abstracting and generalizing from past experiences. Previous research focused on how humans generalize from isolated sequential processes, yet we know little about mechanisms that facilitate adaptation to more complex dynamics, including those that govern much everyday experience. Here, using a novel sequence learning task based on graph factorization, coupled with simultaneous magnetoencephalography (MEG) recordings, we ask whether reuse of experiential “building blocks” is reflected in neural representations that could support inference and generalization. Behavioral results were consistent with participants decomposing task experience into subprocesses, both abstracting their dynamical structure away from their sensory specifics and reusing these in a new task environment. Neurally, this was accompanied by a learning-induced, condition-specific, increase in neural similarity among stimuli conditional on their shared structural role in abstract subprocesses across distinct task phases. Consistent with this, a complementary dynamical role decoding analysis revealed enhanced neural decodability, across experiential contexts, for stimuli that adhered to identical structural transition types, specifically when prior experience included the relevant graph structure. Decoding strength for these role representations was positively related to behavioral success in reuse of subprocess knowledge, a relationship that will require future independent replication. The findings provide neural evidence consistent with the notion that abstract role representations may contribute to the reuse of experience within new contexts.