Research Associate University of Maryland Baltimore School of Medicine
Abstract Text: Introduction One of the most common symptoms in mTBI patients is working memory impairment which may persist years post-mTBI [1]. The N-back fMRI is likely the most popular measures of working memory. Despite growing literature on the use of N-back test among TBI patients, no clear picture has emerged between the performance on N-back test and the corresponding functional activation patterns in mild TBI. A clear understanding of this relationship may provide better characterization of mTBI subjects particularly when no cognitive differences are observed between the mTBI patients and control subjects. In this study, we investigated various N-back task loads effects on the functional activation pattern in patients with mTBI compared with Healthy Controls (HCs), as well as longitudinal changes within the mTBI group.
Methods Thirty-five patients with mTBI (46.46±15.89 years old) and 28 age-matched HCs were included. Cognitive test and N-back fMRI scans were performed for mTBI patients at post-1month and post-6month injury and once for HC subjects. Military Acute Concussion Evaluation (MACE) and Modified Rivermead Post Concussive Questionnaire (RPQ) were collected for mTBI patients [2]. Three functional runs (0-, 1-, and 2-back) were acquired in order using a T2-weighted EPI sequence (TE/TR = 30 /2300 ms, resolution = 2.396 x 2.396 mm2, slice thickness = 4mm). Each functional run consisted of 5 task blocks. Each block lasted 48.3 seconds and was followed by 23 seconds of visualizing “+” sign. The task block consisted of the presentation of a sequence of letters and participants responded with a MR-compatible response box. Accuracy and average reaction time for correct responses were recorded. The fMRI data were processed including slice timing, realignment, normalization to the MNI template, and spatially smoothed with 8-mm FWHM. Statistical analyses were performed on individual data using the GLM, while group analyses were performed using a random-effects model. The significant level was defined as voxel-wise uncorrected p < 0.008 and cluster-wise FDR-corrected p < 0.05 for multiple comparison.
Results No significant differences were observed in age, gender, education level, and MACE between the mTBI and HC and in Rivermead score across time in mTBI. We observed no significant difference in accuracy between the HCs and mTBI groups for 0-back and 1-back tasks, however, significant difference were observed for the more demanding 2-back task. No significance was observed in mean reaction time. For all the three task loads, consistent task positive brain regions in fronto-parietal network, cerebellum, and basal ganglia network, as well as task negative brain regions in default mode network (DMN) among mTBI and HCs was observed. For mTBI at post-1month, higher activation was found in left parietal when comparing 1-back with 0-back, in bilateral fronto-parietal network when comparing 2-back with 1-back load, and in bilateral fronto-parietal network and bilateral cerebellum when comparing 1-back with 0-back. For mTBI at post-6month, significant activation difference was observed only between 2-back and 0-back tasks with higher activation within the fronto-parietal network, and lower activation within the DMN. No significant activation difference was observed within HCs. Significant difference between HC controls and mTBI were observed post-1month and post-6month respectively for the 2-back task. In both these comparisons increased activation were observed within the fronto-parietal network. No significant longitudinal activation changes within the mTBI patients for all the three task loads.
Discussion Our results indicate similar accuracy in performance of the N-back task between HCs and mTBI patients, a trend towards increased reaction time in mTBI patients. In terms of functional activities, no differences were found between the HCs and mTBI subjects for the zero and 1-back tests. However, significant differences were observed on the 2-back task where mTBI patients had greater activation within the fronto-parietal network, basal ganglia and cerebellum. The default mode network was also suppressed only for the 2-back tests for mTBI patients. Our results suggested with appropriate working memory load, it would be able to detect subtle functional changes in mTBI patients, even though differences in behavioral performance between mTBI and HCs were absent.
Keywords: Traumatic brain injury, N-back fMRI, working memory