In this study, we aimed to explore whether MI ability was associated with the cortical activity of brain regions of the MI network in a group of healthy subjects during a gait imagery task. Precisely, results for the vividness of imageries tested with KVIQ and VMIQ questionnaires, and the ability of imagery, measured using mental chronometry assessments were correlated with power changes in β band, recorded with hdEEG.
The main result of this study was that we found a significant relationship between the power of the activity in the areas already shown 37–50 to be involved in MI of gait and the MI ability scores.
Precisely, a positive correlation was detected between VMIQ-ext scores and ERDs β band power of frontal and cingulate areas, and between IA scores and the power of activity of the left inferior frontal and superior temporal regions.
To date, only few studies tried to investigate whether the individual ability to imagine vividly was associated with distinctive brain activity patterns51,53,54. The first paper, published in 1992 by Charlot and co-workers53, measured brain activity in healthy undergraduate volunteers, classified, as "high" and "low" imagers, based on the score of two clinical tests (i.e., the Minnesota Paper Form Board and the Mental Rotations Test), during a visual imagery task consisting in a mental representation and exploration of an imaginary island. Using regional cerebral blood flow (CBF) imaging, they found that low imagers had a widespread CBF increase, whereas high imagers showed a more focal activation53. This finding was explained by authors hypothesizing a low cognitive functions differentiation in bad imagers and, conversely, a more differentiated cognitive architecture in skilled imagers. Later, differences in brain activity, using functional magnetic resonance imaging (fMRI), were investigated in participants showing high or poor MI ability, during both physical execution and mental imagery of finger movements54. Results revealed that good imagers had a higher bilateral activation in the premotor, parietal regions, known to have crucial role in the MI network, with respect to bad imagers. By contrast, participants with poor MI ability manifested greater posterior cingulate, orbitofrontal areas, and cerebellum activations, possibly reflecting a compensatory mechanism to counteract difficulties in creating a vivid representation of sequential movements.
Concerning evidence investigating differences among subjects with good and poor MI ability during MI of gait, it has been reported that imagery capacities do influence functional brain activity even during the imagination of a simple and well automatized motor task. Meulen and collaborators51, in fact, found that participants with good MI ability had a higher cortical activation in the primary motor cortex, the prefrontal cortex, thalamus, and cerebellum with respect to those with lower imagery performance.
In line with these results, also here we found that MI ability level influenced cortical recruitment specifically in those areas which are particularly involved in the MI neural network. Precisely, a positive correlation was found between the MI ability test scores and the left inferior and middle frontal areas, the precentral regions, and the SMAs, suggesting that the better the IA, the more the involvement of these areas. Frontal activity is known to be crucial for MI and especially for MI of gait, supporting the fact that gait is no longer considered a simple and automatic motor action. Indeed, various cognitive functions (such as attention and visuo-spatial abilities) are involved during walking, especially during complex tasks, thus justifying the intervention of frontal regions for being in charge of the higher-order cognitive control of gait55.
Temporal areas were involved through the left superior temporal region activation, showing a positive correlation with IA SCORE. Temporal regions are usually recognized to participate in allocentric processing, fundamental for activities involving spatial memory and navigation. Even if also egocentric processing is implicated in navigation, allocentric processing might help in maintaining a cognitive representation of the environment by updating our own location within it and in avoiding cumulative errors associated with egocentric representation56. This assumption might explain the higher activation of temporal regions in participants who had a better objective ability in performing MI, measured via mental chronometry.
Finally, scores obtained by volunteers during the third person VMIQ test, were significantly positively correlated with avgERD(i) of middle cingulum. This brain region is recognized to be part of the MI network, and its activity results to be crucial for performing MI, specifically when considering MI of usual gait52.
It is worthy to note that no significant correlations were found between brain activity and scores of the first-person visual perspective of KVIQ (i.e., KVIQ-v) and of the of VMIQ (i.e., VMIQ-int). This could be related to the nature of our gait imagery task, where the external strategy might fit better when observing a path and imagining of moving forward.
A possible explanation might be represented by the different brain processes that took place when subjects have to execute visual MI in first person respect to a third-person perspective. Indeed, it was recently speculated that first-person imagery uses a bottom-up strategy, thus taking into account actions and reactions to concrete aspects of the imagined environment, whereas third-person imagery uses a top-down strategy due to the integration of the MI event with its wider context, including experience of other events beyond the main one57.
Finally, no results revealed a negative correlation between MI ability tests and MI network activity, supporting the hypothesis that a finest MI ability is associated with a higher recruitment of regions involved in MI network.
Several limitations of the study deserve attention. First, the small sample size lessened the strength of our results. Second, leg muscle activity was not recorded during the hdEEG registration. Nonetheless, previous studies showed that EMG activity of distal leg muscles recorded during seated position decreased, while during standing gait MI tasks led to a facilitatory effect on proximal lower limb muscle activity58,59. According to these findings we might suppose an irrelevant effect of leg muscle activity on EEG data acquisition. Third, even though cerebellar activity has been linked to MI cortical network, data consistency of hdEEG in detecting signals from the cerebellum is still up for debate18.