Event-Related Potentials as Markers of Efficacy for Combined Working Memory Training and Transcranial Direct Current Stimulation Regimens: A Proof-of-Concept Study

Abstract

Our brains are often under pressure to process a continuous flow of information in a short time, therefore facing a constantly increasing demand for cognitive resources. Recent studies have highlighted that a lasting improvement of cognitive functions may be achieved by exploiting plasticity, i.e., the brain's ability to adapt to the ever-changing cognitive demands imposed by the environment. Transcranial direct current stimulation (tDCS), when combined with cognitive training, can promote plasticity, amplify training gains and their maintenance over time. The availability of low-cost wearable devices has made these approaches more feasible, albeit the effectiveness of combined training regimens is still unclear. To quantify the effectiveness of such protocols, many researchers have focused on behavioral measures such as accuracy or reaction time. These variables only return a global, non-specific picture of the underlying cognitive process. Electrophysiology instead has the finer grained resolution required to shed new light on the time course of the events underpinning processes critical to cognitive control, and if and how these processes are modulated by concurrent tDCS. To the best of our knowledge, research in this direction is still very limited. We investigate the electrophysiological correlates of combined 3-day working memory training and non-invasive brain stimulation in young adults. We focus on event-related potentials (ERPs), instead of other features such as oscillations or connectivity, because components can be measured on as little as one electrode. ERP components are, therefore, well suited for use with home devices, usually equipped with a limited number of recording channels. We consider short-, mid-, and long-latency components typically elicited by working memory tasks and assess if and how the amplitude of these components are modulated by the combined training regimen. We found no significant effects of tDCS either behaviorally or in brain activity, as measured by ERPs. We concluded that either tDCS was ineffective (because of the specific protocol or the sample under consideration, i.e., young adults) or brain-related changes, if present, were too subtle. Therefore, we suggest that other measures of brain activity may be more appropriate/sensitive to training- and/or tDCS-induced modulations, such as network connectivity, especially in young adults.

Keywords: electroencephalography; electrophysiological markers; event related potential; non-invasive brain stimulation; plasticity; transcranial direct current stimulation; working memory training; young adults.

FIGURE 1

Diagram describing the timeline of the procedure.

FIGURE 2

Exemplification of tasks and stimuli used in this study. Above, change detection task; below, n-back task.

FIGURE 3

Region of interest and corresponding electrode assignment.

FIGURE 4

Time-course of the ROIs selected to measure ERP components. Below, topographies of the components considered, with the arrow linking them to the corresponding peak on the ROIs time course.

FIGURE 5

Diagram explain strategy instructions: participants were instructed to create a memory array with the first “n” items in the stream, then to create a second array with the next “n” items. At this point they could compare “new” items with “old” ones and respond. Finally, they had to discard the “old” array, not necessary, and create a new one. The process was repeated until the end of the stream.

FIGURE 6

Changes in average N in the ASNBACK task (training) as a function of TIME and STIMULATION.

FIGURE 7

Changes in reaction time (A), performance (B), capacity (C), and bias (D) in the Change detection task. P-values are indicated as follows: °p < 0.1, *p < 0.05, **p < 0.01, and ***p < 0.001.

FIGURE 8

Changes in reaction time (A), performance (B) and bias (C) at post-test (T4) and follow up (T5) in the SNBACK task. P-values are indicated as follows: *p < 0.05, **p < 0.01, and ***p < 0.001.

FIGURE 9

Amplitude of ERP for each LOAD (n = 2, n = 3, and n = 4), TIME (T0, T4, and T5) and component.

FIGURE 10

Amplitude and modulation changes at post-test (T4) and follow-up (T5), for each component, as measured during the SNBACK task. P-values are indicated as follows: °p < 0.1, *p < 0.05, **p < 0.01.