Kohler Natalie, Czepiel Anna M, de Manzano Örjan, Novembre Giacomo, Keller Peter E, Villringer Arno, Sammler Daniela
Department of Neurology, Max Planck Institute for Human Cognitive and Brain Sciences, Leipzig, Germany.
Research Group Neurocognition of Music and Language, Max Planck Institute for Empirical Aesthetics, Frankfurt am Main, Germany.
Front Hum Neurosci. 2025 Mar 12;19:1543131. doi: 10.3389/fnhum.2025.1543131. eCollection 2025.
During ensemble performance, musicians predict their own and their partners' action outcomes to smoothly coordinate in real time. The neural auditory-motor system is thought to contribute to these predictions by running internal forward models that simulate self- and other-produced actions slightly ahead of time. What remains elusive, however, is whether and how own and partner actions can be represented and in the sensorimotor system, and whether these representations are . Here, we applied multivariate pattern analysis (MVPA) to functional magnetic resonance imaging (fMRI) data of duetting pianists to dissociate the neural representation of self- and other-produced actions during synchronous joint music performance. Expert pianists played familiar right-hand melodies in a 3 T MR-scanner, in duet with a partner who played the corresponding left-hand basslines in an adjacent room. In half of the pieces, pianists were motorically familiar (or unfamiliar) with their partner's left-hand part. MVPA was applied in primary motor and premotor cortices (M1, PMC), cerebellum, and planum temporale of both hemispheres to classify which piece was performed. Classification accuracies were higher in left than right M1, reflecting the content-specific neural representation of self-produced right-hand melodies. Notably, PMC showed the opposite lateralization, with higher accuracies in the right than left hemisphere, likely reflecting the content-specific neural representation of other-produced left-hand basslines. Direct physiological support for the representational alignment of partners' M1 and PMC should be gained in future studies using novel tools like interbrain representational similarity analyses. Surprisingly, motor representations in PMC were similarly precise irrespective of familiarity with the partner's part. This suggests that expert pianists may generalize contents of familiar actions to unfamiliar pieces with similar musical structure, based on the auditory perception of the partner's part. Overall, these findings support the notion of parallel, distinct, and content-specific self and other internal forward models that are integrated within cortico-cerebellar auditory-motor networks to support smooth coordination in musical ensemble performance and possibly other forms of social interaction.
在合奏表演过程中,音乐家会预测自己和同伴的动作结果,以便实时顺利地进行协调。神经听觉 - 运动系统被认为通过运行内部前向模型来做出这些预测,这些模型会稍微提前模拟自我和他人产生的动作。然而,尚不清楚的是,自我和同伴的动作能否以及如何在感觉运动系统中得到表征,以及这些表征是否……在此,我们将多变量模式分析(MVPA)应用于二重奏钢琴家的功能磁共振成像(fMRI)数据,以区分同步联合音乐表演过程中自我和他人产生动作的神经表征。专业钢琴家在3T磁共振扫描仪中弹奏熟悉的右手旋律,与在相邻房间弹奏相应左手低音线的同伴进行二重奏。在一半的曲目演奏中,钢琴家对同伴的左手部分在运动上熟悉(或不熟悉)。MVPA应用于两个半球的初级运动皮层和运动前皮层(M1、PMC)、小脑和颞平面,以分类演奏的是哪首曲目。左M1的分类准确率高于右M1,反映了自我产生的右手旋律的内容特异性神经表征。值得注意的是,PMC表现出相反的偏侧化,右半球的准确率高于左半球,这可能反映了他人产生的左手低音线的内容特异性神经表征。未来使用脑间表征相似性分析等新工具的研究应能获得对同伴M1和PMC表征对齐的直接生理支持。令人惊讶的是,无论对同伴部分是否熟悉,PMC中的运动表征都同样精确。这表明专业钢琴家可能会根据对同伴部分的听觉感知,将熟悉动作的内容推广到具有相似音乐结构的不熟悉曲目中。总体而言,这些发现支持了并行、不同且内容特定性的自我和他人内部前向模型的概念,这些模型整合在皮质 - 小脑听觉 - 运动网络中,以支持音乐合奏表演以及可能其他形式的社会互动中的顺利协调。