A longitudinal proton magnetic resonance spectroscopy study of mild traumatic brain injury

Abstract

Despite the prevalence and impact of mild traumatic brain injury (mTBI), common clinical assessment methods for mTBI have insufficient sensitivity and specificity. Moreover, few researchers have attempted to document underlying changes in physiology as a function of recovery from mTBI. Proton magnetic resonance spectroscopy (¹H-MRS) was used to assess neurometabolite concentrations in a supraventricular tissue slab in 30 individuals with semi-acute mTBI, and 30 sex-, age-, and education-matched controls. No significant group differences were evident on traditional measures of attention, memory, working memory, processing speed, and executive skills, though the mTBI group reported significantly more somatic, cognitive, and emotional symptoms. At a mean of 13 days post-injury, white matter concentrations of creatine (Cre) and phosphocreatine (PCre) and the combined glutamate-glutamine signal (Glx) were elevated in the mTBI group, while gray matter concentrations of Glx were reduced. Partial normalization of these three neurometabolites and N-acetyl aspartate occurred in the early days post-injury, during the semi-acute period of recovery. In addition, 17 mTBI patients (57%) returned for a follow-up evaluation (mean = 120 days post-injury). A significant group × time interaction indicated recovery in the mTBI group for gray matter Glx, and trends toward recovery in white matter Cre and Glx. An estimate of premorbid intelligence predicted the magnitude of neurometabolite normalization over the follow-up interval for the mTBI group, indicating that biological factors underlying intelligence may also be associated with more rapid recovery.

FIG. 1.

Location of the supraventricular axial proton magnetic resonance spectroscopy imaging (¹H-MRSI) slice in the sagittal and axial planes.

FIG. 2.

Graphical depiction of mean metabolite levels at visit 1 for patients with mild traumatic brain injury (mTBI; black bars) compared to healthy controls (gray bars). Error bars are equivalent to the standard deviation, and all values are corrected for premorbid intelligence. (A) Graph illustrating creatine (Cre) and glutamate-glutamine (Glx) in white matter (WM). (B) Graph depicting Glx in gray matter (GM). Units for the graphs are represented in millimoles per kilogram (mmol/kg) water. Two asterisks (**) denote a significant result (p < 0.05).

FIG. 3.

Graphical illustration of mean white matter (WM) and gray matter (GM) N-acetyl aspartate (A, NAA) and choline (B, Cho) levels at visit 1 for patients with mild traumatic brain injury (mTBI; black bars) compared to healthy controls (gray bars). Error bars are equivalent to the standard deviation and all values are corrected for premorbid intelligence. Units for the graphs are represented in millimoles per kilogram (mmol/kg) water.

FIG. 4.

Graphical representation of mean metabolite change scores (visit 2 – visit 1) for patients with mild traumatic brain injury (mTBI; black bars) compared to healthy controls (gray bars). Error bars are equivalent to the standard deviation and all values are corrected for premorbid intelligence. (A) Graph depicting creatine (Cre) in white matter (WM). (B) Graph depicting glutamate-glutamine (Glx) in WM. (C) Graph depicting Glx in gray matter (GM). Units for the graphs are represented in changes in millimoles per kilogram (mmol/kg) water. A single asterisk (*) denotes a statistical trend, (.05 ≤ p < .10) whereas two asterisk (**) denote a significant effect (p < .05).