The role of individual variability on the predictive performance of machine learning applied to large bio-logging datasets.

Animal-borne tagging (bio-logging) generates large and complex datasets. In particular, accelerometer tags, which provide information on behaviour and energy expenditure of wild animals, produce high-resolution multi-dimensional data, and can be challenging to analyse. We tested the performance of commonly used artificial intelligence tools on datasets of increasing volume and dimensionality. By collecting bio-logging data across several sampling seasons, datasets are inherently characterized by inter-individual variability. Such information should be considered when predicting behaviour. We integrated both unsupervised and supervised machine learning approaches to predict behaviours in two penguin species. The classified behaviours obtained from the unsupervised approach Expectation Maximisation were used to train the supervised approach Random Forest. We assessed agreement between the approaches, the performance of Random Forest on unknown data and the implications for the calculation of energy expenditure. Consideration of behavioural variability resulted in high agreement (> 80%) in behavioural classifications and minimal differences in energy expenditure estimates. However, some outliers with < 70% of agreement, highlighted how behaviours characterized by signal similarity are confused. We advise the broad bio-logging community, approaching these large datasets, to be cautious when upscaling predictions, as this might lead to less accurate estimates of behaviour and energy expenditure.

The information score as a measure of oral discourse comprehension in the early school years.

Purpose: Oral discourse comprehension is a fundamental skill that is predictive of future reading comprehension. The ability to identify important semantic propositions, in a narrative text, and their attachment to a mental model is an essential component of discourse comprehension. This study aimed to pilot a measure of free-recall in a narrative retell across three age groups to explore development. A secondary aim compared measures of free-recall and cued-recall.Method: One hundred and thirty-two (132) typically developing children aged 4-7 years provided a retell of the Squirrel Story narrative and completed the Narrative Comprehension Assessment (NCA). The Squirrel Story Information Score (SSIS) was used as a novel measure of free-recall calculated from the narrative retells. The NCA, which involves literal and inferential comprehension questions, was used as a measure of cued-recall. Scoring reliability was calculated to ensure test-retest validity.Result: Correlation analysis found excellent reliability of the SSIS measure. The difference between 4 and 5 year olds, and 5 and 6 year olds, was significant, with a large effect size. The difference between the 5 and 6 year olds was not significant. A large positive correlation was found between the measures of SSIS and NCA.Conclusion: The measure of SSIS proved sensitive to developmental progression, with a positive relationship found between the measures of free-recall and cued-recall. Variability between the measures highlights the need to consider both measures to ensure an accurate representation of discourse comprehension skills.

Personal and couple level risk factors: Maternal and paternal parent-child aggression risk.

Previous literature examining parent-child aggression (PCA) risk has relied heavily upon mothers, limiting our understanding of paternal risk factors. Moreover, the extent to which factors in the couple relationship work in tandem with personal vulnerabilities to impact PCA risk is unclear. The current study examined whether personal stress and distress predicted PCA risk (child abuse potential, over-reactive discipline style, harsh discipline practices) for fathers as well as mothers and whether couple functioning mediated versus moderated the relation between personal stress and PCA risk in a sample of 81 couples. Additionally, the potential for risk factors in one partner to cross over and affect their partner's PCA risk was considered. Findings indicated higher personal stress predicted elevated maternal and paternal PCA risk. Better couple functioning did not moderate this relationship but partially mediated stress and PCA risk for both mothers and fathers. In addition, maternal stress evidenced a cross-over effect, wherein mothers' personal stress linked to fathers' couple functioning. Findings support the role of stress and couple functioning in maternal and paternal PCA risk, including potential cross-over effects that warrant further inquiry.

Multimethod assessment of children's distress during noninvasive outpatient medical procedures: child and parent attitudes and factors.

The present study assessed behavioral distress during noninvasive outpatient procedures in children ages 4-10 years using a multimethod approach. Factors hypothesized to exacerbate children's distress included parents' and children's attitudes toward healthcare, children's knowledge of medical settings, and parental anxiety. A total of 53 parent-child dyads were recruited from outpatient clinics, with procedural distress assessed via child report, parent report, and direct observation. Some differences emerged depending on the method used to gauge distress. Children's healthcare attitudes and knowledge were associated with observed and child-reported distress, but parent's personal anxiety was associated only with their own perceptions of children's procedural distress. Parents' attitudes toward healthcare were associated with their anxiety but not with children's behavioral distress or healthcare attitudes. Findings are discussed in terms of more consistent findings regarding children's healthcare knowledge and attitudes versus the potential need for additional research on more divergent findings regarding parents' anxiety and healthcare attitudes.

Behind the cycle of violence, beyond abuse history: a brief report on the association of parental attachment to physical child abuse potential.

Although the concept of a cycle of violence presumes that the transmission of violence is expressed directly across generations, the role of the overall quality of the parent-child relationship may ultimately be more influential in later parenting behavior. This study investigated whether mothers' poorer attachment to their parents was associated with their current increased child abuse potential and dysfunctional disciplinary style independent of a personal history of child abuse. A sample of 73 at-risk mothers raising children with behavior problems reported on their parental attachment, abuse potential, dysfunctional parenting style, and personal abuse history. An at-risk sample, rather than a sample of identified abuse victims or perpetrators, was studied to better examine the potential continuity or discontinuity from history of abuse to current abuse risk, allowing consideration of those who may break the cycle versus those who potentially initiate abuse in the absence of a personal history. Findings indicate that poor attachment significantly predicted both dysfunctional parenting practices and elevated child abuse potential, controlling for personal child abuse history. Such results highlight the importance of the overall quality of the relationship between the parent and child in potentially shaping future abuse risk. Findings are discussed in terms of continuity or discontinuity in the cycle of violence and future directions for research on attachment in relation to the development of later child abuse risk.