IRUS Total

Causality and emergence in complex systems: Perspectives from information theory

File Description SizeFormat 
Rajpal-H-2021-PhD-Thesis.pdfThesis68.51 MBAdobe PDFView/Open
Title: Causality and emergence in complex systems: Perspectives from information theory
Authors: Rajpal, Hardik
Item Type: Thesis or dissertation
Abstract: Information theory is becoming increasingly useful in the field of complex systems. In part because of the greater availability of data from various real-world systems but especially because of its ability to capture complex interdependencies in systems with many interacting parts. This thesis explores many such applications of information theory on a wide range of systems to understand two central themes in complexity sciences: \textit{Causality} and \textit{Emergence}. Inspired by Judea Pearl's ladder of causation, the type of analysis conducted in order to achieve the goals of this thesis are divided into three main parts, \textit{Associations}, \textit{Interventions} and \textit{Modelling}. These are the three probes used in this thesis to study causality in complex systems in this thesis. The first part deals with passively observing local, network-level and higher-order differences in the statistical associations in the brain activity that lead to an altered resting-state perception in schizophrenia. Here we find two interesting differences, firstly an increase in directed functional connectivity from frontal to the posterior brain regions and abnormal higher-order information processing in the parietal region of the brain. In the second part, we study the improvisational state of mind in musicians by studying their brain activity. We also explore the impact of this intervention on the audience that witnesses this improvisation in a concert-experiment setting. The findings here relate closely to the entropic brain hypothesis and yield useful and objective measures of engagement among the audience. Finally, we explore models of perception, opinion dynamics and evolution in the later part of this thesis that uses mathematical modelling to explore emergent phenomena like psychosis, polarization and species-environment interactions. These models enable us to highlight the role of aberrant Bayesian priors in schizophrenia and psychedelics, echo chambers in polarization and higher-level organization of species in ecosystems.
Content Version: Open Access
Issue Date: Jul-2021
Date Awarded: Oct-2021
URI: http://hdl.handle.net/10044/1/92841
DOI: https://doi.org/10.25560/92841
Copyright Statement: Creative Commons Attribution-NonCommercial 4.0 International Licence
Supervisor: Jensen, Henrik Jeldtoft
Rosas De Andraca, Fernando
Sponsor/Funder: Imperial College, London
Department: Mathematics
Publisher: Imperial College London
Qualification Level: Doctoral
Qualification Name: Doctor of Philosophy (PhD)
Appears in Collections:Mathematics PhD theses

This item is licensed under a Creative Commons License Creative Commons