Repeated preand post-synaptic neuronal activation is fundamental for strengthening synaptic connections, a key mechanism referred to as spike-time-dependent plasticity (STDP) [1]. In humans, associative plasticity with STDP properties can be induced through a TMS protocol, named cortico-cortical paired associative stimulation (ccPAS) [2e4]. By administering repeated pairs of TMS pulses over two interconnected brain areas at specific interstimulus intervals (ISI), ccPAS allows for the modulation of cortico-cortical connections efficiency. To date ccPAS has been predominantly applied to corticocortical motor pathways [2e4]. For example, following ventral premotor-to-motor cortex (PMv-to-M1) ccPAS, scholars documented a strengthening of the targeted circuit, indexed by the increase of the (inhibitory) effect of PMv conditioning over ipsilateral M1 excitability at rest [2] and the increase in restingstate connectivity of the broader functional network encompassing PMv-M1 areas [3]. Effects of increased connectivity are long-lasting [2,4], anatomically specific [2,3] and associated with functionally specific behavioral gains [4]. All the aforementioned studies reported plastic effects induced by ccPAS when the selected ISI met the temporal rules of shortlatency (supposedly direct) connections, informed by dual-site TMS (dsTMS) [5]. Notably, recent dsTMS studies tested the chronometry of PMv-to-M1 interactions and showed that they occur at different time scales [5e7]. For example, conditioning PMv was found to reduce the size of motor-evoked potentials (MEPs) induced by stimulation of ipsilateral M1 not only at a 8-ms ISI (short-latency interaction) [5], but also at longer (e.g., 40-ms) ISIs [6], thus demonstrating long-latency, likely indirect, inhibitory PMv-to-M1 interactions. Despite this notion, there is no evidence that ccPAS protocols based on long-latency interactions (i.e., ll-ccPAS) can induce associative plasticity in humans. Here we empirically address this question by testing the effect of 3 ll-ccPAS protocols on PMv-M1 interactions in healthy volunteers (see Supplementary information for details on methods). In the PMv-to-M1 ll-ccPAS group (N 1⁄4 12), we continuously administered 90 pairs of TMS pulses over the left PMv and the left M1 at a rate of 0.1 Hz [2e4]. For each pair, PMv preceded M1 stimulation by 40 ms. Such ISI was aimed at activating long-latency PMv-to-M1 inhibitory connections [6]. To test for neuroanatomical specificity [2], we administered the same llccPAS protocol over a parallel pathway connecting the supplementary motor areas (SMA) to M1 (i.e., SMA-to-M1 ll-ccPAS; N 1⁄4 12).

中文翻译：

---

通过基于长延迟皮质 - 皮质相互作用的新型配对联想刺激驱动运动前 - 运动连接中的联想可塑性

重复的突触前和突触后神经元激活是加强突触连接的基础，这是一种称为尖峰时间依赖性可塑性 (STDP) [1] 的关键机制。在人类中，可以通过 TMS 协议诱导具有 STDP 特性的关联可塑性，称为皮质-皮质配对关联刺激 (ccPAS) [2e4]。通过以特定的刺激间隔 (ISI) 在两个相互连接的大脑区域上管理重复的 TMS 脉冲对，ccPAS 允许调节皮质-皮质连接效率。迄今为止，ccPAS 主要应用于皮质皮质运动通路 [2e4]。例如，在腹侧运动前到运动皮层 (PMv-to-M1) ccPAS 之后，学者们记录了目标回路的强化，由 PMv 调节对同侧 M1 静息兴奋性的（抑制性）效应增加和包含 PMv-M1 区域的更广泛功能网络的静息状态连接性增加 [3] 索引。增加连接的影响是持久的 [2,4]、解剖学特定的 [2,3] 并与功能特定的行为增益相关 [4]。所有上述研究报告了CCPA诱导的塑料效应当所选ISI满足缺点（据说直接）连接时的时间规则，由双站点TMS（DSTMS）[5]通知。值得注意的是，最近的 dsTMS 研究测试了 PMv 与 M1 相互作用的计时，并表明它们发生在不同的时间尺度 [5e7]。例如，发现调节 PMv 可减少刺激同侧 M1 引起的运动诱发电位 (MEP) 的大小，不仅在 8 毫秒 ISI（短延迟交互）[5]，而且在更长的时间（例如，40 毫秒） ) ISI [6]，因此证明了长延迟、可能是间接的、抑制性 PMv 与 M1 的相互作用。尽管有这个概念，但没有证据表明基于长延迟交互（即 ll-ccPAS）的 ccPAS 协议可以诱导人类的联想可塑性。在这里，我们通过测试 3 ll-ccPAS 协议对健康志愿者 PMv-M1 相互作用的影响来凭经验解决这个问题（有关方法的详细信息，请参见补充信息）。在 PMv 到 M1 ll-ccPAS 组 (N 1⁄4 12) 中，我们以 0.1 Hz 的速率在左侧 PMv 和左侧 M1 上连续管理 90 对 TMS 脉冲 [2e4]。对于每一对，PMv 在 M1 刺激之前 40 毫秒。这种 ISI 旨在激活长延迟 PMv-to-M1 抑制连接 [6]。为了测试神经解剖学的特异性 [2]，我们通过将辅助运动区 (SMA) 连接到 M1（即 SMA-to-M1 ll-ccPAS；N 1⁄4 12）的平行通路实施了相同的 llccPAS 协议。