Functional connectivity between the thalamus and visual cortex under eyes closed and eyes open conditions: a resting-state fMRI study

Abstract

The thalamus and visual cortex are two key components associated with the alpha power of electroencephalography. However, their functional relationship remains to be elucidated. Here, we employ resting-state functional MRI to investigate the temporal correlations of spontaneous fluctuations between the thalamus [the whole thalamus and its three largest nuclei (bilateral mediodorsal, ventrolateral and pulvinar nuclei)] and visual cortex under both eyes open and eyes closed conditions. The whole thalamus show negative correlations with the visual cortex and positive correlations with its contralateral counterpart in eyes closed condition, but which are significantly decreased in eyes open condition, consistent with previous findings of electroencephalography desynchronization during eyes open resting state. Furthermore, we find that bilateral thalamic mediodorsal nuclei and bilateral ventrolateral nuclei have remarkably similar connectivity maps, and resemble to those of the whole thalamus, suggesting their crucial contributions to the thalamus-visual correlations. The bilateral pulvinar nuclei are found to show distinct functional connectivity patterns, compatible with previous findings of the asymmetry of anatomical and functional organization in the nuclei. Our data provides evidence for the associations of intrinsic spontaneous neuronal activity between the thalamus and visual cortex under different resting conditions, which might have implications on the understanding of the generation and modulation of the alpha rhythm.

Figure 1

Masks of the thalamus and visual areas and seed thalamic ROIs. (A) Mask covering the thalamus and visual areas (BA 17, 18, and 19). The thalamus and visual areas was extracted using the automated anatomical labeling (AAL) template and the Brodmann template, respectively, in the MRIcro software. Notably, in the Brodmann template, a small part of the vermis was misclassified into the visual areas. This part was manually delineated and excluded from the current mask. The thalamus was displayed in yellow and the visual cortex in red. ‘L’ denotes the left hemisphere of the brain and ‘R’ denotes the right hemisphere. Z‐axial coordinates in the Talairach and Tournoux space are from −24 to 31 mm in steps of 5 mm. (B) The locations of Pulvinar (‘p’ in the figure), Mediodorsal nucleus (‘dm’ in the figure) and Ventrolateral nucleus (‘vl’ in the figure) were shown in a schematic drawing of the thalamus according to the Talairach and Tournoux atlas (Talairach and Tournoux, 1988). The voxels we selected as seed ROI for P, MD, and VL were marked with plus signs.

Figure 2

Within‐condition and between‐condition functional connectivity maps of the whole thalamus. Within‐condition [eyes closed (EC) and eyes open (EO)] functional connectivity maps of the whole thalamus were shown on the left part. Between‐condition differences of the functional connectivity maps between EC and EO were shown on the right part. Within‐condition and between‐condition maps of left thalamus were shown on the top row, and those of the right thalamus were shown on the bottom row. ‘L’ denotes the left hemisphere of the brain and ‘R’ denotes the right hemisphere. ‘tha’ denotes thalamus. The numbers below the images refer to the z coordinates in the Talairach and Tournoux space. Statistical threshold was set at a significance level of P < 0.05 by combining individual voxel P < 0.05 with a minimum cluster size of 810 mm³ using a Monte Carlo simulation algorithm.

Figure 3

Average correlation coefficients between the thalamus and the visual cortex, and between the bilateral thalamus. (A) Within‐ and between‐condition connectivity values between the thalamus and visual cortex. (B) Within‐ and between‐condition connectivity values between the bilateral thalamus. For EC and EO, the height of each bar represented mean correlation coefficient within each condition, and the error bar represented standard deviation. For EC‐EO, the height of each bar represented the mean value of correlation differences between the conditions, and the error bar represented standard deviation. We also showed the t values and corresponding P values which indicated the significance of the mean values differing from zero. Each circle represented one subject in the figure. ‘L’ and ‘R’ denote seed regions in the left and right hemisphere, respectively. ‘tha’ denotes thalamus and ‘vis’ denotes visual cortex.

Figure 4

Within‐condition and between‐condition functional connectivity maps of the six thalamic nuclei. Within‐condition (EC and EO) functional connectivity maps of the mediodorsal (MD, Fig. 4A), ventrolateral (VL, Fig. 4B) and pulvinar (P, Fig. 4C) nuclei of thalamus were shown on the left part. Between‐condition differences of the functional connectivity maps between EC and EO were shown on the right part. Top rows: within‐condition and between‐condition maps of left seeds, and bottom rows: within‐condition and between‐condition maps of right seeds. ‘L’ denotes the left hemisphere of the brain and ‘R’ denotes the right hemisphere. The numbers below the images refer to the z coordinates in the Talairach and Tournoux space. Statistical threshold was set at a significance level of P < 0.05 by combining individual voxel P < 0.05 with a minimum cluster size of 810 mm³ using a Monte Carlo simulation algorithm. It should be noted that the between‐condition difference maps of left MD, left P, and right VL nuclei were corrected by combining individual voxel P < 0.05 with a minimum cluster size of 405 mm³ due to the less significant results.

Figure 5

Average correlation coefficients between the thalamic nuclei and the visual cortex, and between the thalamic nuclei and the thalamus. (A) Within‐ and between‐condition connectivity values between the MD and visual cortex. (B) Within‐ and between‐condition connectivity values between the MD and the thalamus. (C) Within‐ and between‐condition connectivity values between the VL and visual cortex. (D) Within‐ and between‐condition connectivity values between the VL and the thalamus. (E) Within‐ and between‐condition connectivity values between the P and visual cortex. (F) Within‐ and between‐condition connectivity values between the P and the thalamus. For other details, see Figure 3.

Figure 6

Within‐condition and between‐condition functional connectivity maps (without global signal removal) of the whole thalamus. The between‐condition difference maps of right thalamus were corrected by combining individual voxel P < 0.05 with a minimum cluster size of 405 mm³. For other details, see Figure 2.

Figure 7

Group ICA results under EC. Using ICA, the functional data under EC condition was decomposed into 20 components using Infomax algorithm. We found that thalamus and visual cortex belonged to a single component. The ICA results revealed anticorrelated (negative) relationship between these two regions under EC (|t| > 2.9). Before the ICA analysis, several preprocessing steps [slice timing, motion correction, spatial normalization and spatial smoothing (FWHM = 4 mm)] were performed. While using a similar analysis of the functional data under EO condition, we did not find such a temporal relationship (data not shown).