AI for Psychometrics: Validating Machine Learning Models in Measuring Emotional Intelligence with Eye-Tracking Techniques.

AI, or artificial intelligence, is a technology of creating algorithms and computer systems that mimic human cognitive abilities to perform tasks. Many industries are undergoing revolutions due to the advances and applications of AI technology. The current study explored a burgeoning field-Psychometric AI, which integrates AI methodologies and psychological measurement to not only improve measurement accuracy, efficiency, and effectiveness but also help reduce human bias and increase objectivity in measurement. Specifically, by leveraging unobtrusive eye-tracking sensing techniques and performing 1470 runs with seven different machine-learning classifiers, the current study systematically examined the efficacy of various (ML) models in measuring different facets and measures of the emotional intelligence (EI) construct. Our results revealed an average accuracy ranging from 50-90%, largely depending on the percentile to dichotomize the EI scores. More importantly, our study found that AI algorithms were powerful enough to achieve high accuracy with as little as 5 or 2 s of eye-tracking data. The research also explored the effects of EI facets/measures on ML measurement accuracy and identified many eye-tracking features most predictive of EI scores. Both theoretical and practical implications are discussed.

Somatic Marker Production Deficits do not Explain the Relationship between Psychopathic Traits and Utilitarian Moral Decision Making.

In moral dilemma tasks, high levels of psychopathic traits often predict increased utilitarian responding-specifically, endorsing sacrificing one person to save many. Research suggests that increased arousal (i.e., somatic marker production) underlies lower rates of utilitarian responding during moral dilemmas. Though deficient somatic marker production is characteristic of psychopathy, how this deficit affects the psychopathy-utilitarian connection remains unknown. We assessed psychopathic traits in undergraduates, as well as behavioral performance and skin conductance level reactivity (SCL-R; a measure of somatic marker production) during a moral dilemma task. High psychopathic traits and low SCL-R were associated with increased utilitarian decisions in dilemmas involving direct personal harm. Psychopathic traits were unrelated to SCL-R, nor did SCL-R mediate the relationship between psychopathy and utilitarianism. The present study did not find evidence that somatic marker production explains the connection between utilitarianism and psychopathy in a college population. Further research is necessary to identify the neural mechanisms relating psychopathy and moral decision-making in nonclinical samples.

The clinical impact of copy number variants in inherited bone marrow failure syndromes.

Inherited bone marrow failure syndromes (IBMFSs) comprise a genetically heterogeneous group of diseases with hematopoietic failure and a wide array of physical malformations. Copy number variants (CNVs) were reported in some IBMFSs. It is unclear what impact CNVs play in patients evaluated for a suspected diagnosis of IBMFS. Clinical and genetic data of 323 patients from the Canadian Inherited Marrow Failure Registry from 2001 to 2014, who had a documented genetic work-up, were analyzed. Cases with pathogenic CNVs (at least 1 kilobasepairs) were compared to cases with other mutations. Genotype-phenotype correlations were performed to assess the impact of CNVs. Pathogenic nucleotide-level mutations were found in 157 of 303 tested patients (51.8%). Genome-wide CNV analysis by single nucleotide polymorphism arrays or comparative genomic hybridization arrays revealed pathogenic CNVs in 11 of 67 patients tested (16.4%). In four of these patients, identification of CNV was crucial for establishing the correct diagnosis as their clinical presentation was ambiguous. Eight additional patients were identified to harbor pathogenic CNVs by other methods. Of the 19 patients with pathogenic CNVs, four had compound-heterozygosity of a CNV with a nucleotide-level mutation. Pathogenic CNVs were associated with more extensive non-hematological organ system involvement (p=0.0006), developmental delay (p=0.006) and short stature (p=0.04) compared to nucleotide-level mutations. In conclusion, a significant proportion of patients with IBMFSs harbor pathogenic CNVs which were associated with a more extensive non-hematological phenotype in this cohort. Patients with a phenotype suggestive of IBMFSs but without identification of pathogenic nucleotide-level mutations should undergo specific testing for CNVs.