Repetitive transcranial magnetic stimulation over dorsolateral prefrontal cortex modulates value-based learning during sequential decision-making.

Adaptive behavior in daily life often requires the ability to acquire and represent sequential contingencies between actions and the associated outcomes. Although accumulating evidence implicates the role of dorsolateral prefrontal cortex (dlPFC) in complex value-based learning and decision-making, direct evidence for involvements of this region in integrating information across sequential decision states is still scarce. Using a 3-stage deterministic Markov decision task, here we applied offline, inhibitory low-frequency 1-Hz repetitive transcranial magnetic stimulation (rTMS) over the left dlPFC in young male adults (n = 31, mean age = 23.8 years, SD = 2.5 years) in a within-subject cross-over design to study the roles of this region in influencing value-based sequential decision-making. In two separate sessions, each participant received 1-Hz rTMS stimulation either over the left dlPFC or over the vertex. The results showed that transiently inhibiting the left dlPFC impaired choice accuracy, particularly in situations in which the acquisition of sequential transitions between decision states and temporally lagged action-outcome contingencies played a greater role. Estimating parameters of a diffusion model from behavioral choices, we found that the diffusion drift rate, which reflects the efficiency of information integration, was attenuated by the stimulation. Moreover, the effects of rTMS interacted with session: individuals who could not efficiently integrate information across sequential states in the first session due to disrupted dlPFC function also could not catch up in performance during the second session with those individuals who could learn sequential transitions with intact dlPFC function in the first session. Taken together, our findings suggest that the left dlPFC is crucially involved in the acquisition of complex sequential relations and in the potential of such learning.

Dynamics of fMRI patterns reflect sub-second activation sequences and reveal replay in human visual cortex.

Neural computations are often fast and anatomically localized. Yet, investigating such computations in humans is challenging because non-invasive methods have either high temporal or spatial resolution, but not both. Of particular relevance, fast neural replay is known to occur throughout the brain in a coordinated fashion about which little is known. We develop a multivariate analysis method for functional magnetic resonance imaging that makes it possible to study sequentially activated neural patterns separated by less than 100 ms with precise spatial resolution. Human participants viewed five images individually and sequentially with speeds up to 32 ms between items. Probabilistic pattern classifiers were trained on activation patterns in visual and ventrotemporal cortex during individual image trials. Applied to sequence trials, probabilistic classifier time courses allow the detection of neural representations and their order. Order detection remains possible at speeds up to 32 ms between items (plus 100 ms per item). The frequency spectrum of the sequentiality metric distinguishes between sub- versus supra-second sequences. Importantly, applied to resting-state data our method reveals fast replay of task-related stimuli in visual cortex. This indicates that non-hippocampal replay occurs even after tasks without memory requirements and shows that our method can be used to detect such spontaneously occurring replay.

Replay in minds and machines.

Experience-related brain activity patterns reactivate during sleep, wakeful rest, and brief pauses from active behavior. In parallel, machine learning research has found that experience replay can lead to substantial performance improvements in artificial agents. Together, these lines of research suggest that replay has a variety of computational benefits for decision-making and learning. Here, we provide an overview of putative computational functions of replay as suggested by machine learning and neuroscientific research. We show that replay can lead to faster learning, less forgetting, reorganization or augmentation of experiences, and support planning and generalization. In addition, we highlight the benefits of reactivating abstracted internal representations rather than veridical memories, and discuss how replay could provide a mechanism to build internal representations that improve learning and decision-making.