Imaging of the recovery from stuttering reveals spontaneous neuroplasticity

Background: People who stutter (PWS) overactivate a right-hemispheric network during speech production when compared with controls [1]. After fluency-shaping therapy, these overactivations shift to the left hemisphere as a result of cortical plasticity [2]. A subset of PWS can recover from stuttering even without therapy. To identify the network reflecting 'ideal' compensation and help to orientate future stuttering therapies, we studied the functional substrates of overt speech in this group compared with normal controls and PWS before and after therapy.

Fig. 1

Fig. 2

Methods: Controls, unaided recovered PWS, and persistent PWS before and after fluency-shaping therapy (13 each, all male) were scanned on a 3 T scanner by acquiring 902 volumes of a standard EPI sequence (TR 2000ms). Syntactically balanced sentences were presented visually for three seconds, preceded by an auditory cue, indicating whether the phrase had to be read silently or overtly. Data analysis in SPM2 consisted of a random-effects group analysis (p<0.001, uncorrected) and region of interest analyses. Results: Thanks to noise-induced fluency all subjects produced fluent speech during scanning. PWS activated right frontoparietal areas more than controls only before therapy (Fig. 1A unmasked, Fig. 1B masked with the group's main effect). Recovered PWS activated the right primary auditory cortex and the left opercular BA 47 more than controls (Fig. 1B and 2), the latter activation highlighting the only region to dissociate significantly between recovered and persistent PWS, while the right opercular BA 47 activated in all PWS more than controls (Fig. 2). Discussion. We present dissociable cortical networks in separate groups performing the same task without behavioural difference. Our results reveal a pre-therapeutic right-hemispheric overactivation which attenuates after fluency-shaping therapy, confirming previous results [1]. The relative overactivation of the bilateral primary auditory cortex of both persistent and recovered PWS may be interpreted as a consequence of an altered auditory feedback loop. Because the right opercular BA 47 had earlier been implicated in the self-initiated reduction of stuttering severity of PWS [3], the overactivation of the left-hemispheric pendant in recovered PWS could therefore be interpreted as a functional correlate of stuttering recovery. References: 1. Brown S, et al. (2005) Hum Brain Mapp 25: 105–117. 2. De Nil LF, et al. (2003)J Fluency Disord 28: 357–379. 3. Preibisch C, et al. (2003) Neuroimage 20: 1356–1364.