MRI assessment of cortical thickness and functional activity changes in adolescent girls following three months of practice on a visual-spatial task

Plasticity of gray matter (GM) in adults has been demonstrated before and after practice of both motor and cognitive tasks [1–3]. Evidence points to plasticity effects for both novel learning of a task and for continued practice of a task once learned [4]. A fundamental question is whether GM increases after practice in a specific brain area are concomitant with functional changes in the same area (site specific), and/or whether different areas in the same network show changes. A secondary question is whether GM increases are associated with increased or decreased brain activity. Functional imaging studies show both directions [5–11]. With two exceptions, none of these included structural assessments so any overlap of functional and structural changes could not be determined. One exception [12] reported limited overlap for a mirror reading task; the other [13] reported no overlap for a simple visual-motor task.

We conducted a straightforward test of the site-specific hypothesis. We first examined whether cortical thickness (CT) increased after three months of practicing a visual-spatial game, Tetris. Then, in the same subjects, we determined blood oxygen level dependent (BOLD) changes after practice. Finally, we examined the overlap of the structural and functional results. Since Tetris involves rapid visual-spatial rotation problem-solving and motor co-ordination, we hypothesized GM increases in pre-motor and post-central gyri and parts of the parietal and occipital lobes as shown in previous studies [5, 14–16].

Tetris-induced BOLD signal increases (p < 0.05, FWE corrected) are similar for both visits (Figure 2, top two panels; Additional Files 1, 2, 3 &4). Most are in frontal and parietal lobes including parts of BAs 6, 7, 40, although the follow-up visit showed additional activations in occipital BAs 17, 18, and 19. For both visits, only a few small clusters (<20 voxels each) showed a significantly decreased BOLD signal during task performance. The same comparisons for the control group showed similar results.

Figure 3 shows both CT change and BOLD signal during Tetris play. For baseline and follow-up, overlap was limited to a few small areas of BAs 6, 7, 22, and 19 (arrows in Figure 3, top and middle rows). The areas with decreased activation over the practice period (follow-up minus baseline) showed no overlap with CT increases (bottom row). To increase statistical power, we did additional analyses, first restricting the CT analyses to regions that showed BOLD changes; and second, restricting the BOLD analyses to regions that showed CT changes. Neither analysis identified any new regions of overlap (p < .005, uncorrected).

We found no apparent overlap between structural and functional findings to support the site-specific hypothesis. Interpreting fMRI data has important limitations. For example, the BOLD signal reflects neuronal mass activity. It does not distinguish between excitatory and inhibitory activity, and a marked change in processing strategies involving an anatomical area does not necessitate a change in BOLD [24]. It is possible that practiced-subjects developed a more sophisticated approach, but if this did not invoke novel brain areas, it might not be detectable using fMRI.

The frontal eye fields in left BA 6 showed the most significant CT change, consistent with other imaging studies of visual-spatial tasks. The same area also showed increased activity while the subjects played at baseline, but there was no additional change after practice. Notably, BA 6 is at a critical juncture between "intention" and "action" [25], linking memory of previous responses and the rewards associated with them [26], and is active when motor plans are altered [27]. The left temporal pole (BA 22/38) also showed change. The temporal poles integrate visual, auditory, tactile, and internal physiological information, suggesting a critical role in multimodal perceptual analysis [28]. Moreover, there are lateralizing differences, with right more associated with emotion and socially relevant episodic memories, and left associated with semantic memory[29]. We interpret the temporal pole changes as reflecting increased multisensory integration to support visual, spatial, and tactile input necessary to succeed at Tetris. That greater cortical change was seen in the left temporal pole, suggests Tetris play is processed as a general cognitive puzzle rather than a discrete episodic memory. Whether practiced Tetris play and associated cortical changes generalize to performance changes in other cognitive domains (e.g. working memory, processing speed, spatial reasoning) is not yet determined.

These data indicate that structural and functional brain changes after practice do not necessarily occur in the same location, at least on the levels assessed with these MRI methods.