INTRODUCTION: The architecture of the human thalamus is currently indirectly inferred from human histopathologic and non-human animal studies. Given that movement-related increases in broadband power from electrocorticography recording have been previously shown to identify specific motor cortical subregions involved with a particular movement, we explored whether this phenomenon can also be used to localize motor subregions of thalamus. METHODS: We performed serial human thalamic local field potentials recordings using an electrode moved through the lateral thalamusduring deep brain stimulation surgery in 10 subjects. During the recording at each location, subjects alternated between 5 second epochs of rest and 5 second periods of contralateral hand movement. Power spectra were generated for each epoch using Welch’s overlapping periodogram method (1 second window, 50% overlap). Spectra from all movement and rest epochs were separately averaged. The difference between the power in the 65 to 95 Hz band during movement versus rest was used as a proxy for task-related change in broadband activity. Task-related differences in prominent low-frequency oscillations (patient-specific sub-ranges within 5 to 30 Hz) were analyzed per individual oscillation and in aggregate. Significance was determined using unpaired T tests at a threshold of p = 0.05. RESULTS: We find that recording sites showing statistically significant increases in broadband power during movement tend to tightly cluster in areas predicted to represent motor thalamus, whereas sites showing statistically significant reduction in the power (desynchronization) of low frequency oscillations are more diffusely distributed. This finding is replicated across all subjects. CONCLUSIONS: These results suggest that a task-dependent increase in broadband power is an electrophysiologic marker of focal thalamic activity and may be useful to delineate specific thalamic subregions on an individual subject basis.