Improved characterization of neuronal dynamics through recurrence network analysisIn his dissertation, Narayan Puthanmadam Subramaniyam introduces the recurrence network analysis as a tool for a nonlinear time series analysis. This enables a better understanding of the dynamics of neuronal signals.
The human brain is one of the most complex systems known to mankind. Particularly, the epileptic brain can be considered as a dynamical system exhibiting multi-stable dynamics. In simple terms, the brain of an epileptic patient undergoes transitions between diﬀerent states. The state between the seizures is known as the inter-ictal state, whereas the ictal state represents the seizure period. The post-ictal state occurs after an epileptic seizure and represents an altered state of consciousness. There is also a pre-ictal state, which is the period that leads up to the transition to a seizure. For an epileptic patient, the transition from the inter-ictal state to the ictal state can be rather abrupt and unpredictable, which makes it a debilitating disease and in many cases also results in psychosocial issues with epileptic patients.
The electroencephalogram (EEG) is the most useful tool to aid the diagnosis of epilepsy, a common neurological disorder aﬀecting roughly 1% of the world’s population. The condition is one of the few common clinical problems that still routinely demand an EEG evaluation.
New approaches that have their roots in nonlinear dynamical systems and the chaos theory are required to formally understand the signatures of diﬀerent dynamic states of epilepsy. Understanding and characterizing the dynamical transitions underlying epileptic EEG signals may lead to the timely prevention of (unpredictable) seizures, resulting in better quality of life for epileptic patients.
The recurrence network is a fascinating approach that exploits the patterns of recurrences in the phase space, which is a fundamental property of a dynamical system used to characterize the geometric properties of the (chaotic) dynamics of the system underlying the observed time series. Using graph theoretical measures, one can characterize the topology of such complex networks and, consequently, gain insights to the underlying system dynamics.
Narayan Puthanmadam Subramaniyam’s dissertation shows that, compared to traditional methods, an approach based on a recurrence network analysis is particularly useful in characterizing the diﬀerent dynamic states underlying epileptic EEG signals.
In particular, the dissertation demonstrates that the proposed method based on recurrence networks can capture the complexity in the organization of epileptic EEG data in diﬀerent dynamic states in a more elaborated fashion than other approaches, such as the nonlinear prediction error or the correlation dimension. The results also show that the recurrence network approach analysis can be performed on extremely short window sizes (approx. 500 samples) and are robust to noise.
One of the many results from the analysis shows that the temporal proﬁle of recurrence network measures start showing signs of robust change well before the clinical identiﬁcation of a seizure. This feature represents the exciting possibility of using these measures as biomarkers to detect the seizures in advance.
Public defence of a doctoral dissertation on Friday, 29 January
M.Sc. Narayan Puthanmadam Subramaniyamin’s doctoral dissertation in the ﬁeld of biomedical engineering entitled ‘Recurrence Network Analysis of EEG Signals: A Geometric Approach’ will be publicly examined at the Faculty of Computing and Electrical Engineering of Tampere University of Technology (TUT) in Auditorium S2 in the Sähkötalo building (address: Korkeakoulunkatu 1, Tampere, Finland) at 12 noon on Friday, 29 January 2016.
The opponent will be Professor Jan Casper de Munck (VU University Medical Center, Netherlands). Professor Jari Hyttinen from the Department of Electronics and Communications Engineering of TUT will act as Chairman.
Narayan Puthanmadam Subramaniyamin (32) comes from Kerala, India and works as a Post-doctoral researcher at the Department of Neuroscience and Biomedical Engineering of Aalto University.
The dissertation is available online at: http://URN.fi/URN:ISBN:978-952-15-3694-6