TEST-RETEST PRECISION OF FUNCTIONAL MR IN SENSORY AND MOTOR TASK ACTIVATION

PURPOSE: To determine the test-retest precision of functional MR maps of regions in the brain ''activated'' by sensory, motor, and cognitive tasks. METHODS: Echo-planar images were acquired at 1.5 T in four sub jects during voluntary motor activity involving the thumb and fingers and during tactile stimulation of the palm. Each subject performed the two tasks twice. Functional images of each task were generated at thr ee thresholds. Test-retest precision was calculated in terms of two ra tios: 1) the pixels activated in both iterations of the tasks in propo rtion to the pixels activated by either iteration of the task, and 2) the ratio modified to include first-order neighboring pixels. The firs t is referred to as pixel precision, and the latter as first-order-nei ghbor piuel precision. RESULTS: In each subject, activation from the f irst and second iteration of each task was located in the same region of the same gyrus. Pixel precision was .57 for the two tasks (at a thr eshold of 0.50). First-order-neighbor precision was greater than .80 f or the two tasks at the same threshold. CONCLUSION: High test-retest p recision can be obtained in functional MR.