Bayesian Inference General Procedures for A Single-subject Test Study

Abnormality detection in the identification of a single-subject which deviates from

the majority of the dataset that comes from a control group is a critical problem.

A common approach is to assume that the control group can be characterised in

terms of standard Normal statistics and the detection of single abnormal subject

is in that context. But in many situations the control group can not be described

in terms of Gaussian statistics and the use of standard statistics is inappropriate.

This paper presents a Bayesian Inference General Procedures for A Single-Subject

Test (BIGPAST), designed to mitigate the effects of skewness under the assumption

that the dataset of control group comes from the skewed Student’s t distribution.

BIGPAST operates under the null hypothesis that the single-subject follows the

same distribution as the control group. We assess BIGPAST’s performance against

other methods through a series of simulation studies. The results demonstrate that

BIGPAST is robust against deviations from normality and outperforms the existing

approaches in terms of accuracy. This is because BIGPAST can effectively reduce

model misspecification errors under the skewed Student’s t assumption. We apply

BIGPAST to a MEG dataset consisting of an individual with mild traumatic brain

injury and an age and gender-matched control group, demonstrating its effectiveness

in detecting abnormalities in the single-subject.