Investigating the Effect of Intrinsic Motivation on Alpha Desynchronization Using Sample Entropy

Abstract

The effect of motivation and attention could play an important role in providing personalized learning services and improving learners toward smart education. These effects on brain activity could be quantified by EEG and open the path to analyze the efficiency of services during the learning process. Many studies reported the appearance of EEG alpha desynchronization during the attention period, resulting in better cognitive performance. Motivation was also found to be reflected in EEG. This study investigated the effect of intrinsic motivation on the alpha desynchronization pattern in terms of the complexity of time series data. The sample entropy method was used to quantify the complexity of event-related spectral perturbation (ERSP) of EEG data. We found that when participants can remember the stimulus, ERSP was significantly less complex than when they cannot. However, the effect of intrinsic motivation cannot be defined by using sample entropy directly. ERSP’s main effect showed that motivation affects the complexity of ERSP data; longer continuous alpha desynchronization patterns were found when participants were motivated. Therefore, we introduced an algorithm to identify the longest continuous alpha desynchronization pattern. The method allowed us to understand that intrinsic motivation has an effect on recognition at the frontal and left parietal area directly.

Decoding violated sensory expectations from the auditory cortex of anaesthetised mice: Hierarchical recurrent neural network depicts separate ‘danger’ and ‘safety’ units

“…This is another perspective on the conventional view of entropy as being proportional to disorganisation, given that disorganised (changing) signals potentially contain more information than completely organised (unchanging) signals. Entropy measures have been used successfully to describe biological neural signals (Angsuwatanakul et al, 2020; Phukhachee et al, 2019); thus, this technique is appropriate for characterising the behaviour of artificial neural signals generated by the model.…”

Decoding violated sensory expectations from the auditory cortex of anaesthetized mice: Hierarchical recurrent neural network depicts separate ‘danger’ and ‘safety’ units

“…This is another perspective on the conventional view of entropy as being proportional to disorganisation, given that disorganised (changing) signals potentially contain more information than completely organised (unchanging) signals. Entropy measures have been used successfully to describe biological neural signals (Angsuwatanakul et al, 2020; Phukhachee et al, 2019), thus this technique is appropriate for characterising the behaviour of artificial neural signals generated by the model.…”

Modelling mouse auditory response dynamics along a continuum of consciousness using a deep recurrent neural network

“…Sample entropy [50] was calculated for each ERP time-series of length 120, matching sub-sequences of length 1 with a tolerance of 0.2; these measurements are plotted by condition in figure 2(c). This metric was developed for quantifying pattern complexity in short and noisy biological signals [47], and has been applied extensively to analyse electrophysiological data [35,51,52]; however, alternative metrics, such as signal variance, could equally have served to order the four sets of ERP waveforms.…”