Signal-enhancing switched protocols to study higher-order cognitive tasks with PET.

Abstract

We tested the effectiveness of a switched protocol when it is used to detect signals that result from the study of a higher-order cognitive task with PET. Using language tasks that have been studied extensively in our laboratories, we compared the signal-to-noise ratio (S/N) and statistical significance of the activation signals detected in PET images of regional cerebral blood flow (rCBF), obtained using a standard activation protocol, and of activity concentration, obtained using a switched protocol. Four volunteers were studied with PET while they were performing synonym-generation and word-repetition tasks (activation and baseline tasks, respectively). Each volunteer had three activation/baseline and three baseline/ activation scans. Data for each scan were collected in two frames (60 and 120 sec long). During the first 60 sec, data were collected using a standard activation protocol. Subjects then switched tasks, and acquisition continued for 120 sec. Two images were obtained from each scan: an rCBF image using the first frame and an activity-concentration image using both frames. Images were transformed into Talairach space, subtracted and averaged within and across subjects. Parametric t-statistic images were generated for each protocol, and the magnitude and significance of the activation signals yielded by the two acquisition methods were compared. All the activation foci detected using measurements of rCBF were detected when the switched protocol was used; this protocol, in addition, yielded better S/N values. The cognitive component introduced by task-switching in switched protocols did not yield extra statistically significant foci. In single subjects, the average improvement in the signal significance from regions of activation, at a 95% confidence level, was between 6% and 25%. When scans were averaged across subjects, the switched protocol yielded improvements in signal statistical significance of up to 38%. We present evidence suggesting that switched protocols can be used to study higher-order cognitive tasks and that they yield activation foci with S/N values that are greater than those of equivalent foci detected using an rCBF protocol. Switched protocols appear to be easy to apply to the testing of higher-order cognitive functions. However, the extra cognitive requirement of switching tasks during data acquisition may be a limiting factor when switched protocols are used to study memory processes.