Date of Award


Document Type

Thesis (Master's)

Department or Program

Psychological and Brain Sciences

First Advisor

Alireza Soltani


The brain integrates information from the external world with multiple timescales of neural dynamics. Among the various timescales, intrinsic timescales and timescales for integration of task-relevant signals (e.g. reward-memory timescales) have widely been studied but mainly independently. Here, we explore the distinct and coordinated contributions of different timescales in population-level encoding, using a recent method to simultaneously estimate various timescales of neural dynamics using electrophysiological recordings in the medial prefrontal cortex (mPFC) and the anterior lateral motor cortex (ALC). To identify the characteristics of each timescale, we evaluated the ability of neurons with different timescales in decoding task-relevant signals. The findings suggest that task relevant timescales provide greater contribution to the population-level encoding of corresponding task relevant signals. Strikingly, the contribution was modulated by the average length of intrinsic timescales of the neuronal subpopulation, but not that of task relevant timescales. Moreover, when we examined the direct association between timescales and behaviors, the length of intrinsic timescales was predictable of behavioral consistency. Together, our findings reveal different but coordinated characteristics of task-relevant and intrinsic timescales in the population coding of task-relevant information. Various lengths of task-relevant timescales were equally important for the encoding of the task-relevant signal regardless of their length, probably because the brain collectively processes the heterogeneous lengths of task-relevant memory to effectively compute and represent the task relevant variables. In contrast, long intrinsic timescales contributed more to the population coding compared to the short intrinsic timescales, which suggests the possibility that long intrinsic timescales possess a higher capacity for spreading the information into the neural network that is encoded by task-relevant timescales. These findings provide new insight into how interactions between dynamics of individual neurons contribute to the encoding of information at the population level.