Can Transsynaptic Modulation of Thalamic Nuclei by tACS be an Alternative to DBS?
DOI: Pending DOI
Synopsis
Alternating current (AC) stimulation applied to the cerebellar surface can entrain neural activity through mono- and di-synaptic connections. Purkinje cells (PCs) exhibited phase-locked modulation and the cerebellar nuclear (CN) cells were entrained through inhibitory PC projections trans-synaptically. The motor thalamic (Mthal) cells were modulated by CN projections disynaptically. Population-level phase locking values (PLV) showed strongest modulation at middle frequencies (~100 Hz), with weaker responses at lower and higher frequencies. Analysis of preferred phase (PPh) of the response revealed increasing propagation delays across successive synaptic layers. These findings demonstrate that cerebellar transcranial AC stimulation (tACS) using surface electrodes can produce frequency-dependent multisynaptic modulation of deep brain structures and may offer an alternative to the invasive deep brain stimulation (DBS) method for some applications.
Background
Electric fields (E-fields) produced by non-invasive stimulation techniques attenuate quickly by depth in the brain, making it hard to modulate subcortical regions without exposing the superficial areas to extreme E-fields [1]. Focal stimulation in deeper brain areas away from the stimulation source is also very difficult with traditional stimulation methods. On the other hand, excitation of distant cells through multi-synaptic pathways using trains of current pulses is a standard procedure in neuroscience experiments. Usually a train of current pulses at a high frequency is employed in these experiments because in most cases a single pre-synaptic spike cannot lead to activation of the post-synaptic cell, and going through multiple synapses is even more challenging.
Interestingly, our experience in the cerebellum is quite different with alternating current (AC) stimulation. We were able to transsynaptically entrain the CN cells [2-4] and even the projection cells in the motor thalamus through this disynaptic pathway under anesthesia in rats [5]. In many cells, the thalamic cells that did not show detectable connections to the cerebellar cortex by a train of current pulses were entrained with AC stimulation. This suggests that affecting the timing of ongoing spikes, as it is done during AC stimulation, is much easier than producing novel spikes in the post-synaptic cell. This opens the possibility of effectively modulating the subcortical areas projected by the neural circuits that are near the brain surface using non-invasive electrodes.
In this paper, we reanalyzed some of the data presented in our prior publications to demonstrate that surface stimulation of the cerebellum may be a potential alternative to DBS for modulation of subcortical brain areas, such as the motor thalamus.
Methods
Details of methodology regarding collection of neural data can be found in earlier publications [3, 5, 6]. In brief, single spikes were recorded from PCs, CN cells, and Mthal cells in ketamine/xylazine or urethane anesthetized rats. Recordings were obtained using glass micropipettes (3-4 MΩ, for PC) and tetrodes (1-2 MΩ, CN & Mthal) inserted perpendicular to the target region through a cranial opening. For PC and CN recordings, AC stimulation was applied on the posterior cerebellum, after removing the skull, using an Ag/AgCl disk electrode positioned on the cerebellar surface and gently pressed against the dura. For Mthal recordings, the skull was not removed [5]. Neural signals were amplified with gains of 100 or 1000 and band-pass filtered between 10 Hz and 10 kHz. Each trial consisted of 40 to 140 seconds of stimulation. The stimulus current amplitude was adjusted to achieve a moderate level of entrainment, corresponding to a PLV of approximately 0.5. PLV and PPh were defined earlier [3]. Single cells were isolated using principal component analysis (PCA). All recordings were reanalyzed to characterize the frequency- and phase-dependent entrainment in PCs, CN cells, and Mthal cells for this study.
Because aggregate population responses frequently exhibited bimodal phase distributions, the conventional PLV metric underestimated the true strength of modulation, as two response peaks separated by approximately 180° partially canceled one another. To address this limitation, a second-order phase-locking metric, PLV2, was used [3]. In this approach, spike phases were computed at twice the stimulation frequency, allowing phase-opposed components to align and providing a more accurate estimate of entrainment for bimodal distributions. In this paper, for population-level analysis, spike times from all recorded cells within each population were pooled into a single aggregate spike train. PLV, PLV2, and PPh were then calculated from the combined spike train of the entire population, rather than by averaging metrics obtained from individual cells.
Results
PCs exhibited robust phase-dependent modulation during AC stimulation (Fig. 1A) and separated into two distinct response populations with approximately 180° phase offset within the stimulation cycle. These two populations, referred to as upward and downward PCs, likely arise from differences in the orientation and morphological asymmetry of PC dendritic arbors relative to the applied electric field. The modulation was strongest at intermediate stimulation frequencies, particularly around 100–200 Hz, while lower modulation was observed at both lower and higher frequencies. Despite the frequency-dependent changes in modulation, the phase offset (~180°) between the two PC populations remained preserved across stimulation frequencies, as reported before [3].
The CN cells exhibited clear frequency-dependent modulation with a peak around 80–150 Hz and substantially weaker modulation at higher frequencies (Fig. 1B). Unlike PCs, CN responses frequently exhibited bimodal temporal distributions, consistent with convergence of 180° phase-opposed PC inputs onto individual CN cells. Since multiple PCs project onto each CN cell, combinations of upward and downward PC inputs can produce unimodal or bimodal modulation patterns depending on the relative strength and timing of convergent inhibitory projections. Similar to the CN cells, thalamic responses frequently showed multimodal distributions (Fig. 1C). Although modulation amplitudes were reduced relative to PCs, modulation remained detectable in thalamic responses, particularly at intermediate stimulation frequencies.
Population-level phase locking analysis demonstrated a consistent frequency dependence across all three neural populations (Fig. 1E). Phase-locking values (PLV2), normalized to the maximal response within each population, revealed strongest modulation at intermediate frequencies with reduced entrainment at lower and higher frequencies. While PCs maintained relatively strong modulation at higher frequencies, modulation progressively weakened in CN and thalamic populations. This reduction in high-frequency synchronization in deeper structures is consistent with temporal dispersion introduced across multisynaptic pathways. Small differences in propagation delays between convergent projections can produce substantial phase variability at higher stimulation frequencies, thereby flattening the aggregate population response. As a result, the range of frequencies in which the cerebellar stimulation was most effective in the PCs became narrower in the CN and even narrower in the Mthal (Fig. 1E).
Analysis of PPh further supported a propagation-based mechanism underlying downstream modulation and refuted direct modulation of post-synaptic cells through volume conduction of the stimulation current (Fig. 1F). PPh increased approximately linearly with stimulation frequency for all cell types, suggesting approximately a constant delay. The slope of the phase-frequency plots increased progressively from PCs to CN and Mthal, indicating larger propagation delays in deeper structures.
Together, these findings demonstrate that weak AC stimulation applied to cerebellar cortex can entrain downstream cells through monosynaptic and disynaptic pathways.
Figure 1. Multisynaptic propagation of AC-induced spike entrainment through cerebellothalamic pathway. (A) Spike histograms of 12 PCs from 5 rats during cerebellar AC stimulation at different frequencies. Colored traces represent individual cells and black traces represent population averages in A through C. PC histograms separated into two groups approximately 180° out of phase relative to the stimulation cycle (solid and dash black traces). (B) Spike histograms of 24 CN cells from 8 rats. (C) Spike histograms of 43 thalamic cells from 16 rats. (D) Sagittal view of the rat brain showing recording locations of the PCs (blue), CN cells (red), and Mthal cells (green). (E) Population-level normalized PLV2 as a function of stimulation frequency for all three cell types. (F) PPh as a function of stimulation frequency for all three types of cells. Phase values for each cell type were offset such that the linear line fits meet at zero.
Discussion
In this paper, we reanalyzed and summarized our previously published data to demonstrate that cerebellar tACS can be a potential alternative to DBS for modulation of deep brain structures such as the thalamus. Fundamental drawbacks of transcranial electric stimulation are the poor penetration and localization of the E-field in the brain. Attempts to overcome these drawbacks using different electrode montages and smaller electrodes provide only marginal improvements. Temporal interference (TI) of two slightly different and high AC frequencies, proposed to improve focality in subcortical regions [7], comes with caveats; the cells near the surface are exposed to much higher E-fields and, based on our findings here, at least in the cerebellum the cortical cells may actually be able to follow the high-frequency stimulation waveform, contrary to the proposed TI mechanism. Transsynaptic modulation with tACS may overcome the penetration and focality of stimulation issues by leveraging the anatomical connections. Interference of tACS modulation patterns at the subcortical areas targeted by anatomically converging pathways can also provide more selectivity [8].
A fundamental difference with DBS, however, is that tACS is a subthreshold modulation technique whereas DBS produces novel spikes or suppresses the spikes in the targeted circuits. Subthreshold neuromodulation techniques including tACS has the advantage of not overwriting but simply modulating the spontaneous activity patterns and thus preserving the information content in the cellular activity. On the other hand, DBS is favourable for its ability to eliminate the aberrant firing patterns when the original activity is completely disrupted by a disease or injury. The tACS being a non-invasive approach maybe preferred over DBS for some clinical applications in the future.
Disclosures
Research supported by NINDS/NIH (RF1NS122741)
References
[1] A. S. Asan, S. Gok, and M. Sahin, "Electrical fields induced inside the rat brain with skin, skull, and dural placements of the current injection electrode," PLoS One, vol. 14, no. 1, p. e0203727, 2019, doi: 10.1371/journal.pone.0203727.
[2] B. Avlar, R. Rahman, S. Vendidandi, E. Cetinkaya, A. S. Asan, M. Sahin, and E. J. Lang, "Modulation of cerebellar cortical, cerebellar nuclear and vestibular nuclear activity using alternating electric currents," Front Syst Neurosci, vol. 17, p. 1173738, 2023, doi: 10.3389/fnsys.2023.1173738.
[3] Q. Kang, A. R. Talesh, E. J. Lang, and M. Sahin, "Transsynaptic modulation of cerebellar nuclear cells: theta AC-burst stimulation," J Neural Eng, vol. 21, no. 6, Dec 13 2024, doi: 10.1088/1741-2552/ad9ad1.
[4] D. Mourra, A. M. Cavalieri, M. M. Casey, M. Sahin, and E. J. Lang, "Cerebellar Transcranial AC Stimulation Produces a Frequency-Dependent Bimodal Cerebellar Output Pattern," Cerebellum, vol. 24, no. 1, p. 21, Jan 2 2025, doi: 10.1007/s12311-024-01756-0.
[5] A. R. Talesh, Q. Kang, E. J. Lang, and M. Sahin, "tACS of the Cerebellum and the Motor Cortex Entrains the Spiking Activity of the Cells in Motor Thalamus in a Frequency Dependent Manner," IEEE Trans Neural Syst Rehabil Eng, vol. PP, Dec 17 2025, doi: 10.1109/TNSRE.2025.3644746.
[6] A. S. Asan, E. J. Lang, and M. Sahin, "Entrainment of cerebellar purkinje cells with directional AC electric fields in anesthetized rats," Brain Stimulation, vol. 13, no. 6, pp. 1548-1558, 2020-11-01 2020, doi: 10.1016/j.brs.2020.08.017.
[7] N. Grossman et al., "Noninvasive Deep Brain Stimulation via Temporally Interfering Electric Fields," Cell, vol. 169, no. 6, pp. 1029-1041 e16, Jun 1 2017, doi: 10.1016/j.cell.2017.05.024.
[8] Q. Kang, A. R. Talesh, E. J. Lang, and M. Sahin, "Electrophysiological Interference for Deep Brain Modulation," Annu Int Conf IEEE Eng Med Biol Soc, vol. 2025, pp. 1-6, Jul 2025, doi: 10.1109/EMBC58623.2025.11254703.