Working memory in patients with mild traumatic brain injury: Functional MR imaging analysis

Purpose: To analyze brain activation patterns in response to tests of working memory after a mild traumatic brain injury (MTBI). Materials and Methods: Research ethics committee approval and patient written informed consent were obtained. Brain activation patterns in response to n-back working memory tasks (n = 1, 2, 3) were assessed with functional magnetic resonance (MR) imaging in 20 patients with MTBI within 1 month after their injury and in 18 healthy control subjects. In n-back working memory tasks, participants monitored a series of number stimuli and were to indicate when the presented number was the same as that presented n back previously. Nine (45%) MTBI patients underwent follow-up functional MR imaging studies 6 weeks later. Digit span, a memory test for how many numbers a person can remember in sequence, and continuous performance test (CPT), a test that measures a person's sustained and selective attention and impulsivity, were also performed before functional MR imaging studies and outside the imager for each participant. Clinical data were analyzed by using t and x2 tests. Withingroup, between-group, and initial and follow-up differences of functional MR imaging data were analyzed by using onesample, two-sample, and paired t tests, respectively. Results: Groups were similar for sex (P = .75), years of education (P = .069), digit span (P = .37 for total score), CPT (P = .31, .27, and .43 for omission error, commission error, and hit reaction time, respectively), and accuracy of n-back working memory performance (P = .90, .11, and .39 for one-, two-, and three-back tasks, respectively). Brain activation patterns differed between MTBI patients and controls in response to increasing working memory loads (P < .01, uncorrected). Control subjects maintained their ability to increase activation in the working memory circuitry with each increase in working memory load. In contrast, MTBI patients were impaired in their ability to increase activation in working memory circuitry under both moderate and high working memory load conditions. However, MTBI patients did show cerebral plasticity, as evidenced by more activation in some areas outside and inside the working memory circuitry as compared with control subjects (P < .01, uncorrected). In the 6-week follow-up study, compared with baseline, MTBI patients showed an improvement of activation in response to increasing working memory loads (P < .05, uncorrected). Conclusion: MTBI-induced differences in working memory functional activity were observed even though differences in behavioral performance between MTBI patients and controls were absent, which suggests that this approach may increase sensitivity to MTBI compared with neuropsychological evaluation alone.