Resolving inconsistent effects of tDCS on learning using a homeostatic structural plasticity model

Abstract

Introduction: Transcranial direct current stimulation (tDCS) is increasingly used to modulate motor learning. Current polarity and intensity, electrode montage, and application before or during learning had mixed effects. Both Hebbian and homeostatic plasticity were proposed to account for the observed effects, but the explanatory power of these models is limited. In a previous modeling study, we showed that homeostatic structural plasticity (HSP) model can explain long-lasting after-effects of tDCS and transcranial magnetic stimulation (TMS). The interference between motor learning and tDCS, which are both based on HSP in our model, is a candidate mechanism to resolve complex and seemingly contradictory experimental observations.

Methods: We implemented motor learning and tDCS in a spiking neural network subject to HSP. The anatomical connectivity of the engram induced by motor learning was used to quantify the impact of tDCS on motor learning.

Results: Our modeling results demonstrated that transcranial direct current stimulation applied before learning had weak modulatory effects. It led to a small reduction in connectivity if it was applied uniformly. When applied during learning, targeted anodal stimulation significantly strengthened the engram, while targeted cathodal or uniform stimulation weakened it. Applied after learning, targeted cathodal, but not anodal, tDCS boosted engram connectivity. Strong tDCS would distort the engram structure if not applied in a targeted manner.

Discussion: Our model explained both Hebbian and homeostatic phenomena observed in human tDCS experiments by assuming memory strength positively correlates with engram connectivity. This includes applications with different polarity, intensity, electrode montage, and timing relative to motor learning. The HSP model provides a promising framework for unraveling the dynamic interaction between learning and transcranial DC stimulation.

Keywords: cell assembly; homeostatic structural plasticity; motor learning; spiking neural network; tDCS.

FIGURE 1

Strong motor engram formed during motor learning. (A) Schematic of the M1 neural network, part of which $(10\%)$ was subject to input that models motor learning. (B) Temporal evolution of neural activity and network connectivity $(\mathrm{\Gamma })$ during a typical motor learning episode (purple shading). The purple and dark gray curves in the upper panel represent the firing rates of the motor engram and of the remaining excitatory neurons, respectively. The purple, dark gray, and light gray curves in the lower panel represent the intra-group connectivity of the memory engram, intra-group connectivity of the non-stimulated excitatory neurons, and inter-group connectivity between both populations. (C) The E-E connectivity matrix at the three time points indicated in panel B (dark blue dots), the fourth panel shows a scaling of polarities and intensities motor engram connectivity. The color scale accounts for connectivity (connection probability).

FIGURE 2

Weak cell assembly formed during tDCS. (A) Schematic of the M1 neural network, part of which $(10\%)$ was weakly polarized by the electrical field applied transcranially. (B, C) Temporal evolution of neural activity and network connectivity, when the membrane potential of neurons was depolarized or hyperpolarized during tDCS. (D, E) The E-E connectivity matrix at four time points indicated in panels B and C (dark blue dots). Depolarizing and hyperpolarizing tDCS triggered different processes of synaptic reorganization, but both lead to a similar persistent increase in connectivity among the stimulated neurons.

FIGURE 3

Uniform tDCS generally hinders motor learning. (A) Uniform tDCS was applied to the entire M1 network where the motor engram is embedded. (B, C) Temporal evolution of firing rate and connectivity under three different conditions: applying uniform tDCS immediately before, during, or immediately after learning input. The red and blue curves represent depolarizing and hyperpolarizing DCs, respectively. The yellow curve represents the condition without tDCS. The green and purple shaded areas represent the period of tDCS application (DC) and motor learning (ML), respectively.

FIGURE 4

Targeted tDCS exerts opposite effects when applied at different stages of learning. (A) Targeted tDCS matched with the entire motor engram. (B, C) Temporal evolution of the motor engram’s firing rate and connectivity, when targeted tDCS was applied either immediately before, during, or immediately after the learning input.

FIGURE 5

Motor engram and tDCS stimulated cell assembly interfere in case of overlap. (A) Schematic of the M1 neural network, where the motor learning engram and the tDCS stimulated cell assembly had $50\%$ overlap. (B) Temporal evolution of firing rate and connectivity of excitatory neurons when applying tDCS after learning. The green and purple curves represent neurons that receive only tDCS or only motor learning input, respectively. The orange dashed curve reflects the overlapped half of the engram, and the dark gray curve represents the non-stimulated excitatory neurons. (C) Connectivity matrix at the three time points indicated in panel B. In the right panel of C, the green square indicates connectivity among the assembly of tDCS-stimulated cells, while the purple square labels the motor learning engram. (D, E) Temporal evolution of neural activity and network connectivity, and the resulting connectivity matrix, when tDCS was applied before motor learning.

FIGURE 6

Unfocused tDCS of different intensities applied before, during, or after learning triggered different results. (A–C) Temporal evolution of firing rate and connectivity of the population that only sees motor learning input, the overlapped part of the motor engram that was also stimulated, and the neurons only subjected to tDCS, respectively.

FIGURE 7

Differential effects of strong tDCS on the motor engram connectivity pattern. Network connectivity was captured at $1200\mathrm{s}$ when the network dynamics reached its equilibrium state. (A) Strong tDCS (hyperpolarizing or depolarizing) was applied before motor learning. The green square labels the connectivity of the tDCS-stimulated cell assembly, while the purple square labels the connectivity of the motor engram. Strong hyperpolarizing and depolarizing tDCS resulted in a different connectivity pattern as compared to the control case where only motor learning input was applied but no tDCS (left panel). (B) Strong tDCS of both polarities was applied after motor learning. Drastic changes in connectivity pattern were observed in the depolarizing tDCS, where the overlapped engram formed a strong connectivity but almost disconnected with the non-overlapped engram.

FIGURE 8

Modulatory effects of tDCS on motor learning as predicted by our model. We estimated the effects of tDCS by the connectivity of the motor engram $({\mathrm{\Gamma }}_{\text{engram}})$ . The $x-\text{axis}(\mathrm{\Delta }{V}_{\text{m}})$ denotes the polarity and intensity of the applied DC. The horizontal solid lines (light gray) indicate the engram’s baseline connectivity without tDCS. (A) Effects of uniform and targeted tDCS on engram connectivity when applied immediately before, during, or immediately after learning. (B) Connectivity of the overlapped half and the non-overlapped half of the motor engram, when tDCS with different polarities and intensities was applied either before, during, or after learning.