Neurochemical changes within human early blind occipital cortex

Abstract

Early blindness results in occipital cortex neurons responding to a wide range of auditory and tactile stimuli. These changes in tuning properties are accompanied by an extensive reorganization of the occipital cortex that includes alterations in anatomical structure, neurochemical and metabolic pathways. Although it has been established in animal models that neurochemical pathways are heavily affected by early visual deprivation, the effects of blindness on these pathways in humans is still not well characterized. Here, using (1)H magnetic resonance spectroscopy in nine early blind and normally sighted subjects, we find that early blindness is associated with higher levels of creatine, choline and myo-Inositol and indications of lower levels of GABA within the occipital cortex. These results suggest that the cross-modal responses associated with early blindness may, at least in part, be driven by changes within occipital biochemical pathways.

Keywords: ANOVA; BOLD; CSF; EB; FID; FWHM; GM; MRS; N-acetyl aspartate; N-methyl-d-aspartate; NAA; NMDA; PET; PRESS; Position Resolved Spectroscopy; S/N; SC; WM; analysis of variance; blindness; blood-oxygen-level-dependent; cerebral spinal fluid; cross-modal plasticity; early blind; free-induction decay; full width at half maximum; gray matter; magnetic resonance spectroscopy; occipital; positron emission tomography; sighted control; signal to noise ratio; visual deprivation; white matter.

Figure 1

(A) The locations of the three MRS voxels shown in a representative early blind individual. A high resolution T1 weighted anatomical scan (MPRAGE: TR = 7 ms; TE = 3.20 ms; flip angle = 8°; matrix size 240×240; 160 sagittal slices; slice thickness = 1 mm) was collected at the beginning of each session to guide voxel placement for the MRS sequences. 3×2×3 cm voxels were placed within visual cortex, along the left and right occipital wall centered around the calcarine sulcus and within a control frontal lobe region, anterior to the central sulcus and superior to the corpus callosum. Care was taken to ensure that voxels did not overlap with the pial surface as to avoid lipid contamination. (B) The short echo (TE =32 ms) spectrum, after model fitting and quantification procedures, showing the locations of several hydrogen MRS peaks: myo-Inositol (mI), choline (Cho), creatine (Cr) and N-acetyl aspartate (NAA). The smooth line below the spectrum is the LCmodel computed spectral baseline (subtracted from the resonance area under the curve when calculating chemical concentrations). (C) The long echo (TE=87ms) spectrum used to quantify GABA and glutamate. Longer echo times allow for the isolation of the glutamate peak due to dephasing of the glutamine signal at this echo time (Puts et al., 2012). The smooth line below the spectrum is the LCmodel computed spectral baseline. (D) The GABA peak was isolated with a frequency-selective saturation MEGAPRESS technique with subtraction across the two different frequency positions of the editing pulse. The thicker line shows the spectrum with the editing pulse positioned at 1.9 ppm with both creatine and choline visible. The thinner line shows the MEGAPRESS spectrum after subtraction of the two editing pulses (7.5 − 1.9 ppm), with GABA visible at 2.95 p.p.m.

Figure 2

Group median tissue concentrations for sighted control (light gray bars) and early blind subjects (dark gray bars). Error bars represent the 25–75% interquartile range.

Figure 3

Early visual deprivation has significant effects on choline, creatine and myo-Inositol and GABA concentrations in occipital but not precentral cortex. Absolute concentrations (in millimolar, mM) are shown across left and right occipital voxels and the precentral gyrus for A) choline, B) creatine, C) myo-Inositol, D) NAA. Group median metabolite concentrations are shown for sighted control (light gray bars) and early blind subjects (dark gray bars). Error bars represent the 25–75% interquartile range. * p < 0.05, ** p<0.01, *** p < 0.001 represent group differences using post-hoc Students t-tests corrected for 6 multiple comparisons using the Bonferroni-Holm method.