Time-Frequency Brown Bag Seminar

Wednesday, March 10, 1999

12:30pm

EQuad E415

Speaker: Jonathan Cohen, Princeton University

Title: Studies of the fMRI "impulse-response" function in the cognitive domain.

Abstract:

Previous research has begun to exploit the temporal resolution of fMRI to functionally dissociate brain regions based on the dynamics of observed activity. For example, in a recent study (Cohen, et al., 1997, Nature) we found sustained load-sensitive activity in dorsolateral prefrontal cortex during working memory (WM) performance. However, a more complex pattern was observed in Broca's area, showing transient increases in signal with both increased amplitude and more prolonged duration for higher levels of memory load. Without better knowledge of the response dynamics of the fMRI signal in these regions, we cannot determine whether such activation dynamics reflect an underlying process that a) lasts the same amount of time but is more intensely engaged by higher memory loads; or b) is equally engaged but lasts longer with higher loads. The current study was performed to help characterize the impulse-response function for the fMRI signal in prefrontal regions, similar to what has been done in sensory regions (Boynton et al., 1996; Vazquez & Noll, 1996). Subjects performed a Sternberg memory task, viewing a briefly-flashed memory set of letters and, after a retention interval, making a yes-no response to a single-letter memory probe to indicate whether it was in the memory set. We manipulated the duration of WM processes by varying the retention interval between 2 and 10 seconds; we manipulated the intensity of WM processes by varying the memory set-size between 2 and 5 letters. Functional scans covering the frontal cortex were obtained every 2 seconds. Some regions within the frontal cortex show a near-linear impulse-response function, but others showed more complex patterns. These findings, and other issues concerning the use of fMRI to study the dynamics of brain activaiton will be discussed.
back to TFBB page