Can we enhance training-induced plasticity by modulating inhibitory cortical circuits with transcranial stimulation? (Commentary on Mix et al.)


Non-invasive methods of brain stimulation in humans such as repetitive transcranial magnetic stimulation (rTMS) or transcranial direct current stimulation (TDCS) have emerged as powerful research tools to study and manipulate cortical plasticity in the human brain (Ziemann et al., 2008). Both rTMS and TDCS have been successfully used to augment training-induced plasticity (Reis et al., 2008). This feature is of particular interest for current attempts to use transcranial brain stimulation therapeutically, for instance to facilitate post-stroke recovery (Bolognini et al., 2009). One prominent strategy is to optimize training-induced plastic changes by priming the cortical area involved in learning. Indeed there is evidence that this priming approach can enhance training-related plasticity and boost learning in humans, yet only a small number of studies have addressed the underlying neurobiological mechanisms in animal studies (e.g. Fritsch et al., 2010). While current research focuses on changes in synaptic and intrinsic excitability in excitatory glutamatergic neurons, the conditioning effects of transcranial stimulation techniques on the activity of inhibitory interneurons has received little attention.

In the article published in this issue of EJN, Annika Mix and collaborators (Mix et al., 2010) addressed the question how transcranial stimulation and subsequent learning influence the cortical expression of proteins linked to inhibitory neurotransmission. They applied theta burst stimulation (TBS) before rats learned a tactile discrimination task, and studied how TBS and learning effects impact on regional protein expression in cortical areas directly involved in the task and cortical control areas that are not subserving the task. TBS has been introduced in 2005 as a highly effective rTMS protocol for inducing bi-directional changes in motor cortex excitability in healthy human volunteers (Huang et al., 2005). Intermittent TBS (iTBS) tends to induce lasting increases in cortical excitability, whereas continuous TBS (cTBS) tends to decrease it (Huang et al., 2005).

In accordance with their recent findings (Trippe et al., 2009), Mix et al. showed that iTBS, but not cTBS, strongly reduced the cortical expression of the calcium-binding protein parvalbumin (PV), which is mainly localized in fast-spiking inhibitory cortical interneurons, accompanied by a decrease in the cytosolic 67-kD isoform of glutamic acid decarboxylase (GAD67). Extending their previous work, they observed that only iTBS, but not cTBS, improved learning success. Moreover, subsequent learning partially prevented the suppressive effect of intermittent TBS on regional PV and GAD67 expression in task-relevant cortical areas such as the barrel cortex, but not in task-irrelevant areas such as the visual cortex.

A different pattern of effects was observed regarding the cortical expression of the 65-kD isoform of GAD65, which is localized in the synaptic terminals. Cortical GAD65 expression was markedly increased by cTBS or learning alone (and moderately by iTBS), but the combination of TBS and learning did not significantly change the effects on GAD65 expression that were induced by each intervention alone.

Together, the present findings represent a critical step towards a better mechanistic understanding of the plastic effects that can be induced with transcranial stimulation in the cortex. The results have two important implications for the use of transcranial stimulation techniques in humans in a therapeutic setting. First, the results clearly indicate that TBS alone has multiple effects on cortical inhibition that to a large extent depend on the specific TBS protocol. Second, the results lend support to the concept that a TBS-induced disinhibition of cortical networks might render cortical circuits more susceptible to training-dependent plasticity, and thereby increase the efficacy of subsequent training to induce faster or stronger behavioural improvement (Siebner, 2010). Electrophysiological studies on motor cortex excitability in humans have shown that the currently available transcranial stimulation protocols such as TDCS, TBS, continuous rTMS or paired associative stimulation differ in terms of their conditioning effects on intracortical inhibition, but share the capability to induce bi-directional changes in corticospinal excitability (Ziemann et al., 2008). These differences in the conditioning effects on intracortical inhibition might be of relevance when brain stimulation protocols are combined with training.

A major difference between the studies in rodents and humans is that in rodents, TBS causes more widespread brain stimulation. Indeed, changes in protein expression were present in frontal, motor, barrel and visual cortex. Nevertheless, changes in protein expression related to improved learning could be specifically induced in the task-related cortical areas. In humans, TBS-induced changes are more focal with the conditioning effects being mainly confined to the stimulated cortex and connected brain regions. Despite some inherent problems with translating the findings from transcranial stimulation studies in rodents to humans, the present study should prompt clinical neuroscientists to focus more on how transcranial stimulation alters cortical inhibition and how a lasting modulation of cortical inhibition promotes training-related plasticity in humans.