Search Abstracts | Symposia | Slide Sessions | Poster Sessions
Comparing tDCS montages and their effects on speech motor learning
Poster Session D, Saturday, September 13, 5:00 - 6:30 pm, Field House
Maria Cuervo Cano1, Hung-Shao Cheng1,2, Errin Checkley1,3, Poppy Fane De Salis1,3, Andrea Leone-Thide1, Adam Buchwald1; 1New York University, 2University of Wisconsin at Madison, 3University of Bath
Introduction. Learning new sound sequences like non-native consonant clusters is a difficult task at both the phonological and motor levels. Previous evidence from our lab suggests that applying tDCS over the motor cortex (M1) enhances learning 2 days after training (Buchwald et al., 2019). Here we expanded on these findings by investigating the interplay between stimulation conditions and aspects of speech sound learning. Specifically, we compared stimulation of a region involved in speech motor coordination (C5 in in the 10-20 system) with a region involved in phonological processing (IFG, F5 in the 10-20 system) when paired with a speech motor learning task. We further probed the relationship between stimulation condition by manipulating tDCS polarity. Method. 76 unimpaired native English speakers completed a 2-day protocol. Participants were randomly assigned to one of the five groups: Motor-Anode (anode:C5, cathode:Fp2), Phonological-Anode (anode:F5, cathode:Fp2), Motor-Cathode (cathode:C5, anode:Fp2), Phonological-Cathode (cathode:F5, anode:Fp2), and a Sham group. All active stimulation groups received 1mA current using a 1x1 Soterix tDCS device for 20 minutes. Sham current ramped up to 1mA over 30sec and then back down. tDCS was paired with the practice session of a speech motor learning paradigm that targeted eight non-native onset consonant clusters (/gd/, /pt/, /fm/, /fn/, /vm/, /vn/, /zg/ /zb/). After a short pre-practice session with feedback, participants performed a practice session (~28 minutes) where they were presented with both auditory and orthographic models and asked to produce them as accurately as possible. Approximately half of the practice session occurred during stimulation (e.g., starting ~6-7 minutes after stimulation) and half after stimulation. Learning was evaluated based on changes seen at Retention 1 (R1; ~30 minutes after practice) and Retention 2 (R2; 2 days later). We measured cluster accuracy with acoustically-informed transcription (Buchwald & Cheng, 2023) as well as more fine-grained acoustic measures (burst-to-burst duration in stop-stop clusters). We used mixed-effects models (logistic for accuracy, linear for acoustic measures) with group*session interaction and random-intercepts for word and participant to evaluate change. Results and Discussion. All groups significantly improved cluster production accuracy from baseline to R(etention)1 and R2 (all ps>.01). The improvement for the Motor-Anode group at R2 was ~2% greater than other groups but the interactions were not significant. For cluster duration, only sham, Motor-Anode and Phonological-Cathode groups showed significant decreases in duration reflecting a more cluster-like production (all ps>.01). This indicates that the conditions that down-regulate motor regions (Motor-Cathode) and up-regulate phonological regions did not exhibit changes associated with clusters. The Motor-Anode group decreases more than all other groups at R1 (all ps>.01), and more than the Motor-Cathode and Phonological-Anode group at R2 (p>.05). Taken together, these results suggest that measures of accuracy and of consonant cluster timing may tap into different aspects of cluster learning. All stimulation groups improve in accuracy but not in timing. Further, upregulating the motor cortex during training supports all aspects of learning, and particularly along the timing dimension where it leads to significant improvement over sham or down-regulation of the motor region.
Topic Areas: Language Production, Speech Motor Control