Effects of Lithium on Magnetic Resonance Imaging Signal Might Not Preclude Increases in Brain Volume After Chronic Lithium Treatment

Abstract

To the Editor:We read with great interest the thoughtful recent article inwhich Cousins et al. (1) suggested that the structuralbrain changes associated with lithium (Li) treatmentmight not be physical increases in brain volume but rather anartifact caused by shortening of the gray matter (GM) T1relaxation times by Li. We agree that interpretation of clinicalneuroimaging studies investigating the effects of Li on brainstructure is unclear, particularly as to the factors underlyingpotential GM changes. However, although Li treatment mightalter contrast in T1 weighted magnetic resonance (MR) images,we here advance clinical and preclinical evidence why this mightnot preclude a tangible change in brain volume.Specifically, clinical magnetic resonance imaging (MRI) studieshave found Li-associated volumetric changes in subregions of thehippocampus, even in absence of global alterations in brainvolume (2,3). It would be difficult to explain the absence of globalchanges in the presence of regionally constrained variations inGM as an artifact of general shortening of T1 relaxation times byLi (1).In addition, Cousins et al. (1) suggested that the effects of Li onMRI signal would predominantly affect morphometric methods,which heavily depend on tissue type segmentation. It is relevantthat Li might affect the shape of brain regions (4), and a recentmeta-analysis (5) as well as mega-analysis (6) showed largerhippocampal volumes in Li-treated patients relative to healthycontrol subjects in studies using manual volumetry. Although theeffects of T1 shortening on shape analyses or manual volumetryhave not yet been investigated, these techniques are lessdependent on segmentation than voxel-based morphometryand rely on it only to detect the edges of regions of interest.Studies combining neuroimaging and neurocognitive mea-sures provide another line of evidence suggesting that the effectof Li on brain structure might not be an artifact of MRI signalalterations. Shortening of T1 relaxation times would not explainwhy increases in hippocampal volumes after Li treatment corre-lated with improvements in memory performance (7).Treatment with Li is associated with an increase in levels ofbrain N-acetylaspartate, a putative neuronal marker. Althoughsome of the spectroscopic studies used brain water as a referencepeak (8), others used creatine as the internal standard (9) andthus would be less biased by changes in brain water T1 times dueto Li treatment.Cousins et al. (1) observed no changes in white matter (WM)T1 times and no evidence for an “artificial” change in WM volume,due to shortening of GM T1 times (1). These findings do nottherefore account for the global as well as regional increases inWM volume or shape after Li treatment (10,11) or the concomi-tant increases in GM and WM (11).At the pre-clinical level, a recent study demonstrated thatmale rats treated for 8 weeks with clinically relevant concentra-tions of Li showed an increase in whole brain volume ( 4.8%)measured by manual segmentation of T