Quantitative measures of discrimination with application to appointment processes.

Bias and discrimination in appointment processes such as hiring decisions (and analogous selection procedures for performance evaluations, promotions, scholarships, and awards), are quantified statistically via the binomial distribution. These statistical measures are described and an easily used webapp for calculating them is provided. The measures considered include the likelihood that a given number of appointments arose from a fair process and the likelihood that an existing process would give rise to a fair outcome if it were repeated. These methods are illustrated by applying them to sex (including gender) discrimination and racial discrimination in senior appointments in the Australian university sector; both conscious and unconscious biases are included. Significant sex discrimination is found to have existed in the appointments of university chief executives (Vice Chancellors) who were in office in 2018, but with a moderate chance that current processes will yield fair outcomes in the future. However, there is no evidence of strong sex discrimination in the country's eight main research universities for senior appointments (i.e., Faculty Deans and members of their governing Boards or Senates) for those in office as of 2021. However, at the same dates, extreme racial discrimination was implicit in the selection procedures for both Vice Chancellors and senior appointments in all these universities. The University of Sydney's senior appointments were found to have had the most racially biased outcomes among the country's eight main research universities. Significantly, there is negligible statistical likelihood of achieving racially unbiased outcomes in the future in any of the contexts considered, unless the selection procedures are significantly modified.

Statistical and agent-based modelling of the transmissibility of different SARS-CoV-2 variants in England and impact of different interventions.

The English SARS-CoV-2 epidemic has been affected by the emergence of new viral variants such as B.1.177, Alpha and Delta, and changing restrictions. We used statistical models and the agent-based model Covasim, in June 2021, to estimate B.1.177 to be 20% more transmissible than the wild type, Alpha to be 50-80% more transmissible than B.1.177 and Delta to be 65-90% more transmissible than Alpha. Using these estimates in Covasim (calibrated 1 September 2020 to 20 June 2021), in June 2021, we found that due to the high transmissibility of Delta, resurgence in infections driven by the Delta variant would not be prevented, but would be strongly reduced by delaying the relaxation of restrictions by one month and with continued vaccination. This article is part of the theme issue 'Technical challenges of modelling real-life epidemics and examples of overcoming these'.

Modelling the impact of reopening schools in the UK in early 2021 in the presence of the alpha variant and with roll-out of vaccination against SARS-CoV-2.

Following the resurgence of the COVID-19 epidemic in the UK in late 2020 and the emergence of the alpha (also known as B117) variant of the SARS-CoV-2 virus, a third national lockdown was imposed from January 4, 2021. Following the decline of COVID-19 cases over the remainder of January 2021, the question of when and how to reopen schools became an increasingly pressing one in early 2021. This study models the impact of a partial national lockdown with social distancing measures enacted in communities and workplaces under different strategies of reopening schools from March 8, 2021 and compares it to the impact of continual full national lockdown remaining until April 19, 2021. We used our previously published agent-based model, Covasim, to model the emergence of the alpha variant over September 1, 2020 to January 31, 2021 in presence of Test, Trace and Isolate (TTI) strategies. We extended the model to incorporate the impacts of the roll-out of a two-dose vaccine against COVID-19, with 200,000 daily vaccine doses prioritised by age starting with people 75 years or older, assuming vaccination offers a 95% reduction in disease acquisition risk and a 30% reduction in transmission risk. We used the model, calibrated until January 25, 2021, to simulate the impact of a full national lockdown (FNL) with schools closed until April 19, 2021 versus four different partial national lockdown (PNL) scenarios with different elements of schooling open: 1) staggered PNL with primary schools and exam-entry years (years 11 and 13) returning on March 8, 2021 and the rest of the schools years on March 15, 2020; 2) full-return PNL with both primary and secondary schools returning on March 8, 2021; 3) primary-only PNL with primary schools and exam critical years (years 11 and 13) going back only on March 8, 2021 with the rest of the secondary schools back on April 19, 2021 and 4) part-rota PNL with both primary and secondary schools returning on March 8, 2021 with primary schools remaining open continuously but secondary schools on a two-weekly rota-system with years alternating between a fortnight of face-to-face and remote learning until April 19, 2021. Across all scenarios, we projected the number of new daily cases, cumulative deaths and effective reproduction number R until April 30, 2021. Our calibration across different scenarios is consistent with alpha variant being around 60% more transmissible than the wild type. We find that strict social distancing measures, i.e. national lockdowns, were essential in containing the spread of the virus and controlling hospitalisations and deaths during January and February 2021. We estimated that a national lockdown over January and February 2021 would reduce the number of cases by early March to levels similar to those seen in October 2020, with R also falling and remaining below 1 over this period. We estimated that infections would start to increase when schools reopened, but found that if other parts of society remain closed, this resurgence would not be sufficient to bring R above 1. Reopening primary schools and exam critical years only or having primary schools open continuously with secondary schools on rotas was estimated to lead to lower increases in cases and R than if all schools opened. Without an increase in vaccination above the levels seen in January and February, we estimate that R could have increased above 1 following the reopening of society, simulated here from April 19, 2021. Our findings suggest that stringent measures were integral in mitigating the increase in cases and bringing R below 1 over January and February 2021. We found that it was plausible that a PNL with schools partially open from March 8, 2021 and the rest of the society remaining closed until April 19, 2021 would keep R below 1, with some increase evident in infections compared to continual FNL until April 19, 2021. Reopening society in mid-April, without an increase in vaccination levels, could push R above 1 and induce a surge in infections, but the effect of vaccination may be able to control this in future depending on the transmission blocking properties of the vaccines.

Modelling the potential impact of mask use in schools and society on COVID-19 control in the UK.

As the UK reopened after the first wave of the COVID-19 epidemic, crucial questions emerged around the role for ongoing interventions, including test-trace-isolate (TTI) strategies and mandatory masks. Here we assess the importance of masks in secondary schools by evaluating their impact over September 1-October 23, 2020. We show that, assuming TTI levels from August 2020 and no fundamental changes in the virus's transmissibility, adoption of masks in secondary schools would have reduced the predicted size of a second wave, but preventing it would have required 68% or 46% of those with symptoms to seek testing (assuming masks' effective coverage 15% or 30% respectively). With masks in community settings but not secondary schools, the required testing rates increase to 76% and 57%.

Physiology-based ERPs in normal and abnormal states.

Evoked response potentials (ERPs) and other transients are modeled as impulse responses using physiology-based neural field theory (NFT) of the corticothalamic system of neural activity in the human brain that incorporates synaptic and dendritic dynamics, firing response, axonal propagation, and corticocortical and corticothalamic pathways. The properties of model-predicted ERPs are explored throughout the stability zone of the corticothalamic system, and predicted time series and wavelet spectra are also analyzed. This provides a unified treatment of predicted ERPs for both normal and abnormal states within the brain's stability zone, including likely parameters to represent abnormal states of reduced arousal.

Evolutionary algorithm optimization of biological learning parameters in a biomimetic neuroprosthesis.

Biomimetic simulation permits neuroscientists to better understand the complex neuronal dynamics of the brain. Embedding a biomimetic simulation in a closed-loop neuroprosthesis, which can read and write signals from the brain, will permit applications for amelioration of motor, psychiatric, and memory-related brain disorders. Biomimetic neuroprostheses require real-time adaptation to changes in the external environment, thus constituting an example of a dynamic data-driven application system. As model fidelity increases, so does the number of parameters and the complexity of finding appropriate parameter configurations. Instead of adapting synaptic weights via machine learning, we employed major biological learning methods: spike-timing dependent plasticity and reinforcement learning. We optimized the learning metaparameters using evolutionary algorithms, which were implemented in parallel and which used an island model approach to obtain sufficient speed. We employed these methods to train a cortical spiking model to utilize macaque brain activity, indicating a selected target, to drive a virtual musculoskeletal arm with realistic anatomical and biomechanical properties to reach to that target. The optimized system was able to reproduce macaque data from a comparable experimental motor task. These techniques can be used to efficiently tune the parameters of multiscale systems, linking realistic neuronal dynamics to behavior, and thus providing a useful tool for neuroscience and neuroprosthetics.

K-complexes, spindles, and ERPs as impulse responses: unification via neural field theory.

To interrelate K-complexes, spindles, evoked response potentials (ERPs), and spontaneous electroencephalography (EEG) using neural field theory (NFT), physiology-based NFT of the corticothalamic system is used to model cortical excitatory and inhibitory populations and thalamic relay and reticular nuclei. The impulse response function of the model is used to predict the responses to impulses, which are compared with transient waveforms in sleep studies. Fits to empirical data then allow underlying brain physiology to be inferred and compared with other waves. Spontaneous K-complexes, spindles, and other transient waveforms can be reproduced using NFT by treating them as evoked responses to impulsive stimuli with brain parameters appropriate to spontaneous EEG in sleep stage 2. Using this approach, spontaneous K-complexes and sleep spindles can be analyzed using the same single theory as previously been used to account for waking ERPs and other EEG phenomena. As a result, NFT can explain a wide variety of transient waveforms that have only been phenomenologically classified to date. This enables noninvasive fitting to be used to infer underlying physiological parameters. This physiology-based model reproduces the time series of different transient EEG waveforms; it has previously reproduced experimental EEG spectra, and waking ERPs, and many other observations, thereby unifying transient sleep waveforms with these phenomena.

Age trends and sex differences of alpha rhythms including split alpha peaks.

OBJECTIVE: To investigate age trends, sex differences, and splitting of alpha peaks of the EEG spectrum in the healthy population. METHODS: An automated multi-site algorithm was used to parametrize the alpha rhythm in 1498 healthy subjects aged 6-86 years. Alpha peaks identified from multiple electrode sites were organized into clusters of similar frequencies whose sex differences and age trends were investigated. RESULTS: Significant age-related trends were observed for frequency, position, and amplitude of dominant alpha peaks. Occipital sites had alpha clusters of higher average frequency, higher power, and greater presence across the scalp. Frequency and power differences were found between the sexes. CONCLUSION: Observed increases in alpha frequency in children and decreases in the elderly were consistent with those from earlier studies. A large fraction of participants (≈ 44%) showed multiple distinct alpha rhythm thus investigations which only examine the alpha frequency with the highest peak power can produce misleading results. The strong dependence of alpha frequency on age and anterior-posterior position indicates use of a fixed alpha frequency band is insufficient to capture the full characteristics of the alpha rhythm. SIGNIFICANCE: This study establishes alpha rhythm parameter ranges (including power and frequency) in the healthy population, and quantifies the variation in alpha frequency across the scalp. The automated characterization enables objective evaluations of alpha band activities over large samples. These findings are potentially useful in testing theories of alpha generation, where splitting of the alpha rhythm has been theoretically predicted to occur in individuals with large differences in axon length between anterior and posterior corticothalamic loops.

Model-based analysis and quantification of age trends in auditory evoked potentials.

OBJECTIVE: The physiological basis for the changes in auditory evoked potentials (AEPs) during development and aging is currently unknown. This study investigates age- and task-related changes via a mathematical model of neuronal activity, which allows a number of physiological changes to be inferred. METHODS: A quantitative, physiology-based model of activity in cortical and thalamic neurons was used to analyze oddball AEPs recorded from 1498 healthy subjects aged 6-86 years. RESULTS: Differences between standard and target responses can be largely explained by differences in connection strengths between thalamic and cortical neurons. The time it takes signals to travel between the thalamus and cortex decreases during development and increases during aging. Strong age trends are also seen in intracortical and thalamocortical neuronal connection strengths. CONCLUSIONS: Changes in AEP latency can be attributed to changes in the thalamocortical signal propagation time. Large changes in the connection strengths between neuronal populations occur during development, resulting in increased thalamocortical inhibition and decreased thalamocortical excitation. Standard and target parameters are similar in children but diverge during adolescence, due to changes in thalamocortical loop activity. SIGNIFICANCE: Model-based AEP analysis links age-related changes in brain electrophysiology to underlying changes in brain anatomy and physiology, and yields quantitative predictions of several currently unknown physiological and anatomical properties of the brain.

Neurophysiological changes with age probed by inverse modeling of EEG spectra.

OBJECTIVE: To investigate age-associated changes in physiologically-based EEG spectral parameters in the healthy population. METHODS: Eyes-closed EEG spectra of 1498 healthy subjects aged 6-86 years were fitted to a mean-field model of thalamocortical dynamics in a cross-sectional study. Parameters were synaptodendritic rates, cortical wave decay rates, connection strengths (gains), axonal delays for thalamocortical loops, and power normalizations. Age trends were approximated using smooth asymptotically linear functions with a single turning point. We also considered sex differences and relationships between model parameters and traditional quantitative EEG measures. RESULTS: The cross-sectional data suggest that changes tend to be most rapid in childhood, generally leveling off at age 15-20 years. Most gains decrease in magnitude with age, as does power normalization. Axonal and dendritic delays decrease in childhood and then increase. Axonal delays and gains show small but significant sex differences. CONCLUSIONS: Mean-field brain modeling allows interpretation of age-associated EEG trends in terms of physiological processes, including the growth and regression of white matter, influencing axonal delays, and the establishment and pruning of synaptic connections, influencing gains. SIGNIFICANCE: This study demonstrates the feasibility of inverse modeling of EEG spectra as a noninvasive method for investigating large-scale corticothalamic dynamics, and provides a basis for future comparisons.

Physiology-based modeling of cortical auditory evoked potentials.

Evoked potentials are the transient electrical responses caused by changes in the brain following stimuli. This work uses a physiology-based continuum model of neuronal activity in the human brain to calculate theoretical cortical auditory evoked potentials (CAEPs) from the model's linearized response. These are fitted to experimental data, allowing the fitted parameters to be related to brain physiology. This approach yields excellent fits to CAEP data, which can then be compared to fits of EEG spectra. It is shown that the differences between resting eyes-open EEG and standard CAEPs can be explained by changes in the physiology of populations of neurons in corticothalamic pathways, with notable similarities to certain aspects of slow-wave sleep. This pilot study demonstrates the ability of our model-based fitting method to provide information on the underlying physiology of the brain that is not available using standard methods.

Neural mechanisms of ERP change: combining insights from electrophysiology and mathematical modeling.

Using a standardized database of EEG data, recorded during the habituation and oddball paradigms, changes in the auditory event-related potential (ERP) are demonstrated on the time scale of seconds and minutes. Based on previous research and a mathematical model of neural activity, neural mechanisms that could account for these changes are proposed. When the stimulus tones are not relevant to a task, N100 magnitude decreases substantially for the first repetition of a stimulus pattern and increases in response to a variant tone. It is argued these short-term changes are consistent with the hypothesis that there is a refractory period in the neural elements underlying the ERP. In the oddball paradigm, when the stimulus tones are task-relevant, the magnitudes of both N100 and P200 for backgrounds decrease over the entire six-minute recording session. It is argued that these changes are mediated by a decreasing arousal level, and consistent with this, a subject's electrodermal activity (EDA) is shown to reduce over the recording session. By fitting ERPs generated by a biophysical model of neural activity, it is shown that the changes in the background ERPs over the recording session can be reproduced by changing the strength of connections between populations of cortical neurons. For ERPs elicited by infrequent stimuli, there is no corresponding trend in the magnitudes of N100 or P300 components. The effects of stimuli serial order on ERPs are also assessed, showing that the N100 for background ERPs and the N100 and P300 for target ERPs increases as the probability, and expectancy, of receiving a task relevant stimulus increases. Cortical neuromodulation by acetylcholine (ACh) is proposed as a candidate mechanism to mediate the ERP changes associated with attention and arousal.