gms | German Medical Science

Objective: In preoperative language mapping, repetitive transcranial magnetic stimulation (rTMS) is usually accompanied with an assessment of picture naming (PN) task performance. The most frequent rTMS-induced behavioral outcome is a delay of the expected naming response. However, delay ratings are often critisized for being observer-dependent. In this study, we measured the variability and explored potential cut-off’s of delay ratings. Moreover, we aimed to elaborate the differences in delay distribution across different brain regions, comparing sentence onset and item naming.

Methods: Twelve healthy, native German volunteers underwent concurrent rTMS and PN task. Video recordings of the sessions were analysed offline. Two time points were considered: the time at sentence onset (T1) and the time at item naming (T2). In addition to the visual/auditive video assessment by two independent raters providing a qualitative outcome assessment, the exact times of T1 and T2 were manually measured by a third rater using a sound editing software. The stimulation coordinates were exported from the rTMS device and were projected on a standard brain for spatial distribution assessment.

Results: A total of 5396 stimulations were performed across all subjects, of which 48 (1%) resulted in a delay onset response, and 327 (6%) in a delay item response. The average T1 was 716 ± 110 ms for responses qualitatively rated as delayed (versus 416 ± 91 ms for non delayed onset events). The difference T2-T1, however, was 679 ± 159 ms for responses rated as delays (and 477 ± 96 ms for non delayed item responses). As for the spatial distribution of stimulations of each category of delays, the delay onset responses were predominantly elicited by stimulation to a close set of areas, which included the central sulcus and precentral gyrus. In contrast, delayed item responses were observed after stimulation of a larger set of regions, which included the inferior frontal gyrus, medial frontal gyrus and superior temporal gyrus.

Conclusion: Our results suggest that sentence delays are often related to rTMS over (primary and secondary) motor regions, which could hold potential value in differentiating from errors related to other parts of the language networks. Conversely, object naming delays are less often related to motor areas, but rather to “classical” regions within the language network.