Cortico-cortical paired associative stimulation in Alzheimer's disease: preliminary EEG evidence of network-level effects
DOI: Pending DOI
Synopsis
Transcranial magnetic stimulation (TMS) techniques such as cortico-cortical paired associative stimulation (ccPAS) can modulate functional connectivity between interconnected brain regions through spike timing-dependent plasticity (STDP). However, whether this STDP-inducing protocol is effective in Alzheimer's disease (AD) remains unclear. This study investigates whether ccPAS can modulate frontoparietal connectivity in patients with mild-to-moderate AD. Fourteen patients with AD underwent three randomised ccPAS sessions targeting the precuneus (PC) and left dorsolateral prefrontal cortex (DLPFC). Resting-state EEG was recorded before and after each session to assess changes in functional connectivity. Preliminary results revealed a direction-dependent (DLPFC-to-PC) modulation of theta-band connectivity between the two stimulated areas, supporting the feasibility of ccPAS as a network-level intervention in AD.
Background
Synaptic dysfunction is a primary contributor to Alzheimer's disease (AD) pathogenesis alongside amyloid and tau pathology. Spike-timing-dependent plasticity (STDP) mechanisms are impaired from early disease stages [1, 2], and this disruption significantly contributes to cognitive decline. Memory loss in AD has been linked to early degeneration of hippocampal and parietal–frontal circuits, resulting in progressive disruption of large-scale networks supporting episodic memory, attention, and executive functions [3]. Several studies have suggested that rTMS is efficacious in restoring altered long-term potentiation (LTP)-like plasticity in AD [4, 5], supported by animal model studies demonstrating restoration of key synaptic mechanisms following repetitive TMS [6]. However, given the network-level nature of AD pathophysiology, multi-site stimulation protocols targeting functionally connected cortical areas have gained increasing interest. Cortico-cortical paired associative stimulation (ccPAS) is a double-coil TMS protocol based on the Hebbian principle of associative plasticity, in which pairs of pulses are delivered over two interconnected cortical sites at an optimal interstimulus interval, allowing activity elicited at the first site to reach the second immediately before stimulation, thereby enhancing or weakening connectivity between the two regions [7]. Notably, ccPAS targeting frontal and parietal cortices induces direction-dependent STDP-like changes [8], supporting its potential clinical translation. This study investigates whether ccPAS can restore STDP mechanisms between the precuneus (PC) and left dorsolateral prefrontal cortex (DLPFC), potentially promoting network resynchronisation in mild-to-moderate AD.
Methods
Fourteen participants (mean age 74.9 ± 5.2 years) with a clinical and biological diagnosis of Alzheimer's disease (CSF biomarkers of AD pathology or amyloid PET positivity) were recruited at the University Hospital of Siena. Participants were eligible if aged 65–85 years, MMSE >20, native Italian speakers, and on stable pharmacological treatment for at least 6 weeks. Before enrolment, all participants underwent a comprehensive neuropsychological assessment covering global cognitive functioning, episodic memory, attention, and executive functions. Participants completed three ccPAS sessions in a randomised crossover design. Stimulation targeted the precuneus (PC) and left DLPFC using two figure-of-eight coils connected to an STM9000 stimulator (Ates-EBNeuro, Florence, Italy), delivering 150 paired pulses every 5 s (0.2 Hz) at 120% of resting motor threshold. In the forward condition, stimulation of the PC preceded left DLPFC stimulation by 10 ms (ccPAS_10); in the backward condition, the order was reversed using the same interstimulus interval (ccPAS_min10); in the simultaneous condition, pulses were delivered at the same time over the two target regions (ccPAS_0). Before and after each session, participants completed tasks assessing episodic memory (RAVLT, immediate and delayed recall), working memory (N-back), and attention/executive functioning (Trail Making Test, Parts A and B). Resting-state EEG was recorded before and after each session using a 64-channel cap (g.tec Medical Engineering GmbH, Austria) to assess ccPAS-induced changes in functional connectivity.
Results
EEG functional connectivity was assessed using the debiased weighted Phase Lag Index (wPLI), computed across all electrode pairs and five frequency bands (delta: 1–4 Hz; theta: 4.5–8 Hz; alpha: 8.5–13 Hz; beta: 13.5–30 Hz; low gamma: 30.5–50 Hz). Connectivity analyses focused on two a priori regions of interest corresponding to the stimulation targets: Precuneus/PCC (CPz, Pz, CP1, CP2, P1, P2) and left DLPFC (F3, FC3, F5, FC5, AF3, FC1). Post- minus pre-stimulation wPLI changes (ΔwPLI) were computed for each condition and compared across condition pairs using cluster-based permutation statistics (FieldTrip; 1000 randomisations; cluster α = 0.05).
All significant effects were observed exclusively during eyes-closed resting state. Compared to ccPAS_0, ccPAS_min10 produced significantly greater cross-ROI connectivity changes in the delta (p = 0.028; CP1, FC3), theta (cluster 1: p = 0.022, F5, FC5; cluster 2: p = 0.049, P2) and beta bands (p = 0.036; FC5). No significant effects were found for ccPAS_10 versus ccPAS_0 in any frequency band. Critically, the direct comparison between ccPAS_10 and ccPAS_min10 revealed significant negative clusters in the delta (p = 0.022; F3, FC3) and theta bands (p = 0.002; CP1, P1, P2, F3, FC3, AF3, FC1), indicating substantially lower post-stimulation connectivity changes following ccPAS_10 relative to ccPAS_min10. A scalar analysis averaging ΔwPLI across all 36 cross-ROI electrode pairs confirmed the direction of these effects: post-stimulation connectivity change was positive following ccPAS_min10 and negative following ccPAS_10 in both the theta band (mean change: +0.013 vs −0.019; p = 0.003, d = 0.93) and the delta band (p = 0.044, d = 0.58). Regarding behavioural outcomes, no significant effects were observed across conditions, likely reflecting the limited sample size and clinical heterogeneity of this preliminary cohort.
Discussion
These preliminary results provide evidence that ccPAS can modulate frontoparietal connectivity in AD and that its effects are direction dependent. A plausible explanation lies in the impaired propagation of neural activity from the precuneus within the DMN in AD [9]: if signals originating from the precuneus fail to effectively engage distant frontal nodes within the critical timing window, STDP-like plasticity mechanisms may be less efficiently recruited when stimulation is delivered in the forward direction (PC-to-DLPFC). The theta-band specificity of the effect is consistent with previous evidence of impaired theta synchronisation between the precuneus and frontal regions in AD, supporting the view that theta oscillations are particularly sensitive to large-scale network dysfunction affecting memory-related circuits [9]. As enrolment is ongoing, analyses in the complete cohort will assess whether connectivity responses vary across stages of the AD continuum and whether they track progressive network disruption and cognitive performance, potentially identifying stage-specific targets for network-based neuromodulation. Future studies should also investigate individualized inter-stimulus intervals (ISI) tailored to each patient's neurophysiological profile and include additional control conditions. Overall, these preliminary findings support the potential of ccPAS both as a neurophysiological probe of circuit-level dysfunction and as a candidate network-based intervention in AD.
Disclosures
The authors declare that they have no relevant financial interests to disclose.
References
1. Di Lorenzo F, et al. (2018). Impaired spike timing dependent cortico-cortical plasticity in Alzheimer's disease patients. J Alzheimers Dis, 66(3), 983-991.
2. Koch G, et al. (2012). Impaired LTP- but not LTD-like cortical plasticity in Alzheimer's disease patients. J Alzheimers Dis, 31(3), 593-599.
3. Jones DT, et al. (2016). Cascading network failure across the Alzheimer's disease spectrum. Brain, 139(Pt 2), 547–562.
4. Koch G, et al. (2018). Transcranial magnetic stimulation of the precuneus enhances memory and neural activity in prodromal Alzheimer's disease. NeuroImage, 169, 302-311.
5. Koch G, et al. (2022). Precuneus magnetic stimulation for Alzheimer's disease: a randomized, sham-controlled trial. Brain, 145(11), 3776–3786.
6. Tan T, et al. (2013). Low-frequency rTMS reverses Aβ1–42-mediated memory deficits in rats. Exp Gerontol, 48(8), 786-794.
7. Rizzo V, et al. (2009). Paired associative stimulation of left and right human motor cortex shapes interhemispheric motor inhibition based on a Hebbian mechanism. Cereb Cortex, 19(4), 907–915.
8. Santarnecchi E, et al. (2018). Modulation of network-to-network connectivity via spike-timing-dependent noninvasive brain stimulation. Hum Brain Mapp, 39(12), 4870-4883.
9. Maiella M, et al. (2024). Breakdown of TMS evoked EEG signal propagation within the default mode network in Alzheimer's disease. Clin Neurophysiol, 167, 177–188.