Balasundaram Kadirvelu | University of Reading (original) (raw)

Uploads

Papers by Balasundaram Kadirvelu

Research paper thumbnail of Variation in global COVID-19 symptoms by geography and by chronic disease: A global survey using the COVID-19 Symptom Mapper

Research paper thumbnail of Comparative evaluation of Ising couplings, kinetic Ising couplings,and partial correlations in inferring structural connectivity

The problem of inferring the structural connections from the functional connections obtained from... more The problem of inferring the structural connections from the functional connections obtained from the activity of the neuronal networks is one of the major challenges in neuroscience. Studies suggest that maximum entropy based Ising models can discount the effect of indirect interactions and provide good results in inferring the underlying structural connections in neuronal networks. Parameters of the kinetic formulation of the Ising models, kinetic Ising models, have been found to agree well with anatomical connectivity in in silico models of neuronal networks. Following this, Ising and kinetic Ising models have attracted attention in the area of connectivity studies. However, the performance of the Ising couplings and kinetic Ising couplings have not been evaluated in comparison with other functional con- nectivity metrics in the microscopic scale of neuronal networks for a varied set of network conditions and network structures. This thesis sets out to resolve this through a comparative evaluation of the ability of Ising cou- plings and kinetic Ising couplings to unravel the structural connections when compared to the widely used functional connectivity metrics of partial and cross-correlations in in silico networks. The thesis presents the finding that the network correlation level deter- mines the relative performance of the functional connectivity metrics in de- tecting the synaptic connections. At weak levels of network correlation, Ising couplings and kinetic Ising couplings yielded better performance when com- pared to partial and cross-correlations. Whereas at strong levels of network correlation, partial correlations detected more structural links when com- pared to other functional connectivity metrics in this study. This result was consistent across varying firing rates, network sizes, densities and topologies. Along with being directional and applicable in nonstationary cases, kinetic Ising couplings also displayed better performance when compared to Ising couplings. The findings of this thesis serv [...]

Research paper thumbnail of Clustering of patient comorbidities within electronic medical records enables high-precision COVID-19 mortality prediction

We present an explainable AI framework to predict mortality after a positive COVID-19 diagnosis b... more We present an explainable AI framework to predict mortality after a positive COVID-19 diagnosis based solely on data routinely collected in electronic healthcare records (EHRs) obtained prior to diagnosis. We grounded our analysis on the [1/2] Million people UK Biobank and linked NHS COVID-19 records. We developed a method to capture the complexities and large variety of clinical codes present in EHRs, and we show that these have a larger impact on risk than all other patient data but age. We use a form of clustering for natural language processing of the clinical codes, specifically, topic modelling by Latent Dirichlet Allocation (LDA), to generate a succinct digital fingerprint of a patients full secondary care clinical history, i.e. their comorbidities and past interventions. These digital comorbidity fingerprints offer immediately interpretable clinical descriptions that are meaningful, e.g. grouping cardiovascular disorders with common risk factors but also novel groupings that...

Research paper thumbnail of Data-derived wearable digital biomarkers predict Frataxin gene expression levels and longitudinal disease progression in Friedreich’s Ataxia

Friedreich’s ataxia (FA) is a neurodegenerative disease caused by the epigenetic repression of th... more Friedreich’s ataxia (FA) is a neurodegenerative disease caused by the epigenetic repression of the Frataxin gene modulating mitochondrial activity in the brain, which has a diffuse phenotypic impact on patients’ motor behavior. Therefore, with current gold-standard clinical scales, it requires 18–24 month-long clinical trials to determine if disease-modifying therapies are at all beneficial. Our high-performance monitoring approach captures the full-movement kinematics from human subjects using wearable body sensor networks from a cohort of FA patients during their regular clinical visits. We then use artificial intelligence to convert these movement data using universal behavior fingerprints into a digital biomarker of disease state. This enables us to predict two different ‘gold-standard’ clinical scores (SCAFI, SARA) that serve as primary clinical endpoints. Crucially, by performing gene expression analysis on each patient their personal Frataxin gene expression levels were poorl...

Research paper thumbnail of Inferring structural connectivity using Ising couplings in models of neuronal networks

Scientific reports, Jan 15, 2017

Functional connectivity metrics have been widely used to infer the underlying structural connecti... more Functional connectivity metrics have been widely used to infer the underlying structural connectivity in neuronal networks. Maximum entropy based Ising models have been suggested to discount the effect of indirect interactions and give good results in inferring the true anatomical connections. However, no benchmarking is currently available to assess the performance of Ising couplings against other functional connectivity metrics in the microscopic scale of neuronal networks through a wide set of network conditions and network structures. In this paper, we study the performance of the Ising model couplings to infer the synaptic connectivity in in silico networks of neurons and compare its performance against partial and cross-correlations for different correlation levels, firing rates, network sizes, network densities, and topologies. Our results show that the relative performance amongst the three functional connectivity metrics depends primarily on the network correlation levels. ...

Research paper thumbnail of Publisher Correction: Inferring structural connectivity using Ising couplings in models of neuronal networks

Scientific reports, Jan 14, 2018

A correction to this article has been published and is linked from the HTML and PDF versions of t... more A correction to this article has been published and is linked from the HTML and PDF versions of this paper. The error has not been fixed in the paper.

Research paper thumbnail of Covid-19 does not look like what you are looking for: clustering symptoms by nation and multi- morbidities reveal substantial differences to the classical symptom triad

COVID-19 is by convention characterised by a triad of symptoms: cough, fever and loss of taste/sm... more COVID-19 is by convention characterised by a triad of symptoms: cough, fever and loss of taste/smell. The aim of this study was to examine clustering of COVID-19 symptoms based on underlying chronic disease and geographical location. Using a large global symptom survey of 78,299 responders in 190 different countries, we examined symptom profiles in relation to geolocation (grouped by country) and underlying chronic disease (single, co- or multi-morbidities) associated with a positive COVID-19 test result using statistical and machine learning methods to group populations by underlying disease, countries, and symptoms. Taking the responses of 7980 responders with a COVID-19 positive test in the top 5 contributing countries, we find that the most frequently reported symptoms differ across the globe: For example, fatigue 4108(51.5%), headache 3640(45.6%) and loss of smell and taste 3563(44.6%) are the most reported symptoms globally. However, symptom patterns differ by continent; India...

Research paper thumbnail of Variation in global COVID-19 symptoms by geography and by chronic disease: A global survey using the COVID-19 Symptom Mapper

Research paper thumbnail of Comparative evaluation of Ising couplings, kinetic Ising couplings,and partial correlations in inferring structural connectivity

The problem of inferring the structural connections from the functional connections obtained from... more The problem of inferring the structural connections from the functional connections obtained from the activity of the neuronal networks is one of the major challenges in neuroscience. Studies suggest that maximum entropy based Ising models can discount the effect of indirect interactions and provide good results in inferring the underlying structural connections in neuronal networks. Parameters of the kinetic formulation of the Ising models, kinetic Ising models, have been found to agree well with anatomical connectivity in in silico models of neuronal networks. Following this, Ising and kinetic Ising models have attracted attention in the area of connectivity studies. However, the performance of the Ising couplings and kinetic Ising couplings have not been evaluated in comparison with other functional con- nectivity metrics in the microscopic scale of neuronal networks for a varied set of network conditions and network structures. This thesis sets out to resolve this through a comparative evaluation of the ability of Ising cou- plings and kinetic Ising couplings to unravel the structural connections when compared to the widely used functional connectivity metrics of partial and cross-correlations in in silico networks. The thesis presents the finding that the network correlation level deter- mines the relative performance of the functional connectivity metrics in de- tecting the synaptic connections. At weak levels of network correlation, Ising couplings and kinetic Ising couplings yielded better performance when com- pared to partial and cross-correlations. Whereas at strong levels of network correlation, partial correlations detected more structural links when com- pared to other functional connectivity metrics in this study. This result was consistent across varying firing rates, network sizes, densities and topologies. Along with being directional and applicable in nonstationary cases, kinetic Ising couplings also displayed better performance when compared to Ising couplings. The findings of this thesis serv [...]

Research paper thumbnail of Clustering of patient comorbidities within electronic medical records enables high-precision COVID-19 mortality prediction

We present an explainable AI framework to predict mortality after a positive COVID-19 diagnosis b... more We present an explainable AI framework to predict mortality after a positive COVID-19 diagnosis based solely on data routinely collected in electronic healthcare records (EHRs) obtained prior to diagnosis. We grounded our analysis on the [1/2] Million people UK Biobank and linked NHS COVID-19 records. We developed a method to capture the complexities and large variety of clinical codes present in EHRs, and we show that these have a larger impact on risk than all other patient data but age. We use a form of clustering for natural language processing of the clinical codes, specifically, topic modelling by Latent Dirichlet Allocation (LDA), to generate a succinct digital fingerprint of a patients full secondary care clinical history, i.e. their comorbidities and past interventions. These digital comorbidity fingerprints offer immediately interpretable clinical descriptions that are meaningful, e.g. grouping cardiovascular disorders with common risk factors but also novel groupings that...

Research paper thumbnail of Data-derived wearable digital biomarkers predict Frataxin gene expression levels and longitudinal disease progression in Friedreich’s Ataxia

Friedreich’s ataxia (FA) is a neurodegenerative disease caused by the epigenetic repression of th... more Friedreich’s ataxia (FA) is a neurodegenerative disease caused by the epigenetic repression of the Frataxin gene modulating mitochondrial activity in the brain, which has a diffuse phenotypic impact on patients’ motor behavior. Therefore, with current gold-standard clinical scales, it requires 18–24 month-long clinical trials to determine if disease-modifying therapies are at all beneficial. Our high-performance monitoring approach captures the full-movement kinematics from human subjects using wearable body sensor networks from a cohort of FA patients during their regular clinical visits. We then use artificial intelligence to convert these movement data using universal behavior fingerprints into a digital biomarker of disease state. This enables us to predict two different ‘gold-standard’ clinical scores (SCAFI, SARA) that serve as primary clinical endpoints. Crucially, by performing gene expression analysis on each patient their personal Frataxin gene expression levels were poorl...

Research paper thumbnail of Inferring structural connectivity using Ising couplings in models of neuronal networks

Scientific reports, Jan 15, 2017

Functional connectivity metrics have been widely used to infer the underlying structural connecti... more Functional connectivity metrics have been widely used to infer the underlying structural connectivity in neuronal networks. Maximum entropy based Ising models have been suggested to discount the effect of indirect interactions and give good results in inferring the true anatomical connections. However, no benchmarking is currently available to assess the performance of Ising couplings against other functional connectivity metrics in the microscopic scale of neuronal networks through a wide set of network conditions and network structures. In this paper, we study the performance of the Ising model couplings to infer the synaptic connectivity in in silico networks of neurons and compare its performance against partial and cross-correlations for different correlation levels, firing rates, network sizes, network densities, and topologies. Our results show that the relative performance amongst the three functional connectivity metrics depends primarily on the network correlation levels. ...

Research paper thumbnail of Publisher Correction: Inferring structural connectivity using Ising couplings in models of neuronal networks

Scientific reports, Jan 14, 2018

A correction to this article has been published and is linked from the HTML and PDF versions of t... more A correction to this article has been published and is linked from the HTML and PDF versions of this paper. The error has not been fixed in the paper.

Research paper thumbnail of Covid-19 does not look like what you are looking for: clustering symptoms by nation and multi- morbidities reveal substantial differences to the classical symptom triad

COVID-19 is by convention characterised by a triad of symptoms: cough, fever and loss of taste/sm... more COVID-19 is by convention characterised by a triad of symptoms: cough, fever and loss of taste/smell. The aim of this study was to examine clustering of COVID-19 symptoms based on underlying chronic disease and geographical location. Using a large global symptom survey of 78,299 responders in 190 different countries, we examined symptom profiles in relation to geolocation (grouped by country) and underlying chronic disease (single, co- or multi-morbidities) associated with a positive COVID-19 test result using statistical and machine learning methods to group populations by underlying disease, countries, and symptoms. Taking the responses of 7980 responders with a COVID-19 positive test in the top 5 contributing countries, we find that the most frequently reported symptoms differ across the globe: For example, fatigue 4108(51.5%), headache 3640(45.6%) and loss of smell and taste 3563(44.6%) are the most reported symptoms globally. However, symptom patterns differ by continent; India...