Complexity Study in Different States of Consciousness Using Brain Waves

Abstract

Human brain is the prime hub for carrying out all body and mind activities, by relaying information in form of electric waves across millions of neurosynaptic junctions. These electric waves represent the neural signals, which can be broadly classified into different frequency ranges, showing varied levels of dominance for different states of consciousness. In this study, during resting state of 7 voluntary participants, we have extracted and studied the complexity of three neural waves, delta (1 – 4 Hz), theta (4 – 8 Hz) and alpha (8 – 12 Hz) waves, which represent different states of consciousness. Linear techniques were first applied to study parameters like ‘Spectral Power’ and ‘Hemispherical Power Asymmetry’ in all three waves. Keeping in mind the non-linear nature of the extracted neural signals, it seemed more appropriate to use non-linear techniques of analysis, to efficiently capture the intrinsic fluctuation details of the time series. Multifractality and complexity of the different waves were computed and compared by the process of ‘Multifractal Detrended Fluctuation Analysis’ (MFDFA) for resting state. Approximate Entropy (ApEn) was computed to understand the predictability of the fluctuations in the neural waves. Statistical tools like One-way ANOVA, and Mahalanobis Distance were also computed to study the nature of significant difference between the three waves, in terms of each of these above-mentioned properties. All parameters were calculated for both left and right hemispherical sections of frontal, temporal, occipital and parietal lobes. Spectral power in all lobes showed prominently high values for delta waves, and overlapping lower values for theta and alpha waves. Complexity values showed a particular trend in most of the cases, showing highest value in delta, followed by alpha and theta. ApEn in all lobes showed highest values for alpha waves, followed by theta and delta. Power asymmetry and MW asymmetry parameters showed distinctly different behaviour for delta, theta, and alpha waves for most of the lobes. Statistical calculations on the computed parameters revealed different degrees of correlation for delta-theta, alpha-theta, and delta-alpha wave pairs. ANOVA revealed valid significant difference between delta-theta and alpha-theta waves in case of Spectral Power values, and between all three waves in case of ApEn values. Mahalanobis Distance between the three waves showed different trends of values for each of the above-mentioned properties. Results showed, parameters computed by linear methods (Spectral Power, and Power-Asymmetry) failed to reveal prominent distinction between all of the three waves, and certain amount of overlap was observed between computed values of alpha and theta waves in some lobes. Non-linear methods of analysis on the other hand, showed prominent trends of distinction between all three neural waves in all lobes, highlighting better efficiency of non-linear techniques to differentiate neural waves during resting state. The paper presents several new information and interesting points of distinction between delta, theta, alpha waves as revealed from non-linear studies, during resting state of the participants, which are manifestations of different states of consciousness.