Facilitation of inferior frontal cortex by tDCS induces perceptual learning of severely degraded speech

Introduction:

Cochlear implants (CI) offer the possibility to reinstate audition in patients with severe hearing disorders. Due to the limited spectral and temporal resolution of the signal, the auditory input provided by CI is degraded and the ability to adapt to it, thus to reacquire speech comprehension is highly variable and often insufficient. Neurocognitive adaptation to the devices is afforded by perceptual learning. This process can be simulated in healthy individuals by artificially degradation of speech (noise-band vocoding) to test how efficient categorical perceptual sensitivity is generalized, a feature constitutive for perceptual learning. As indicated by a recent functional MRI study, perceptual learning of degraded speech modulates activation in a left-lateralized network, including inferior frontal gyrus (IFG) and inferior parietal cortex (IPC). Here we investigate whether anodal transcranial direct current stimulation (tDCS) is capable of modulating perceptual learning.

Material and methods:

In a sham-controlled, parallel design, 36 volunteers were allocated to 3 intervention groups: tDCSanodal over left IFG, left IPC, or sham stimulation. Participants decided on the match between an acoustically degraded and an undegraded written word by forced same-different choice. Acoustic degradation varied in four noise-vocoding levels (2, 3, 4, 6 bands). 100 minimal pairs were trained on 3 consecutive days and tDCSanodal was applied during the first 20 minutes of each training. Perceptual sensitivity (d') for the trained and an equal number of untrained word pairs was tested before and after training. Increases in d' indicate perceptual learning for untrained, and a combination of associative and perceptual learning for trained pairs.

Results:

For the lowest intelligibility level, perceptual learning occurred only when tDCSanodal was applied over left IFG. For trained pairs, improved d' was seen on all intelligibility levels irrespective of tDCS intervention. Over left IPC, tDCSanodal did not modulate learning but introduced a response bias during training. Volunteers were more liberal to respond “same” potentially indicating enhanced perceptual fusion of degraded auditory with undegraded written input.

Discussion:

Our results supply first evidence that neural facilitation of left IFG is capable of inducing perceptual learning of severely degraded speech: Only when IFG was facilitated during training, perceptual learning occurred for severely degraded speech. Remarkably, facilitation of left IPC did not affect perceptual learning but introduced a decision bias during training, thus playing a modulatory role in the same/different discrimination task. Since noise-vocoding simulates the degraded auditory input supplied by cochlear implants our findings are encouraging to further explore the potential clinical benefits of non-invasive brain stimulation in hearing-impairment and restorative hearing therapy.