The body is represented in a somatotopic framework such that adjacent body parts are represented next to each other in the brain. We utilised the organisation of the somatosensory cortex to study the generalisation pattern of tactile perceptual learning. Perceptual learning refers to the process of long-lasting improvement in the performance of a perceptual task following persistent sensory exposure. In order to test if perceptual learning generalises to neighbouring brain/body areas, 12 participants were trained on a tactile discrimination task on one fingertip (using tactile oriented gratings) over the course of four days. Thresholds for tactile acuity were estimated prior to, and following, the training for the ‘trained’ finger and three additional fingers: ‘adjacent’, ‘homologous’ (the same finger as trained but on the opposite hand) and ‘other’ (which was neither adjacent nor homologous to the trained finger). Identical threshold estimating with no training was also carried out for a control group. Following training, tactile thresholds were improved (as compared to the control group). Importantly, improved performance was not exclusive for the trained finger; it generalised to the adjacent and homologous fingers, but not the other finger. We found that perceptual learning indeed generalises in a way that can be predicted by the topography of the somatosensory cortex, suggesting that sensory experience is not necessary for perceptual learning. These findings may be translated to rehabilitation procedures that train the partially-deprived cortex using similar principles of perceptual learning generalisation, such as following amputation or blindness in adults.
Phantom pain has become an influential example of maladaptive cortical plasticity. According to this model, sensory deprivation following limb amputation allows for intra-regional invasion of neighbouring cortical representations into the former hand area of the primary sensorimotor cortex, which gives rise to pain sensations. Over the years, this model was extended to explain other disorders of pain, motor control and tinnitus, and has inspired rehabilitation strategies. Yet, other research, demonstrating that phantom hand representation is maintained in the sensorimotor system, and that phantom pain can be triggered by bottom-up aberrant inputs, may call this model to question. Using fMRI, we identified the cortical area representing the missing hand in a group of 18 arm amputees. This allowed us to directly study changes in the ‘phantom’ cortex associated with chronic phantom pain, using functional connectivity and voxel-based morphometry. We show that, while loss of sensory input is generally characterized by structural degeneration of the deprived sensorimotor cortex, the experience of persistent pain was associated with preserved intra-regional structure and functional organization. Furthermore, consistent with the dissociative nature of phantom sensations from other sensory experiences, phantom pain is also associated with reduced long-range inter-regional functional connectivity. We propose that this disrupted inter-regional connectivity may be consequential, rather than causal, of the retained yet isolated local representation of phantom pain. We therefore propose that, contrary to the maladaptive model, cortical plasticity occurs when powerful and long-lasting subjective sensory experience, most likely due to peripheral inputs, is decoupled from the external sensory environment.
The rubber hand illusion (RHI) is a multisensory (visual, tactile, proprioceptive) illusion in which participants report body ownership over, mislocalize actual hand position to, and feel touches applied to, the rubber hand. For many years, researchers have used changes in perceived hand position, measured by inter-manual pointing, as a more objective measure of the illusion than verbal reports alone. Despite this reliance, there is little evidence to show that the illusion of hand ownership is directly related to perceived hand position. We developed an adaptive staircase procedure to measure perceived hand position, and tested whether the RHI affected perceived hand position. In two experiments we found a significant illusion of ownership, as well as significant changes in perceived hand position, but these two measures were uncorrelated. In a third experiment using more typical RHI procedures, we again replicated significant illusions of ownership and changes in hand position, but again the measures were uncorrelated. We conclude that viewing and feeling touches applied to a dummy hand results in clear illusions of ownership and changes in hand position, but via independent mechanisms.