Search results
Found 12849 matches for
Credit and blame for AI-generated content: Effects of personalization in four countries.
Generative artificial intelligence (AI) raises ethical questions concerning moral and legal responsibility-specifically, the attributions of credit and blame for AI-generated content. For example, if a human invests minimal skill or effort to produce a beneficial output with an AI tool, can the human still take credit? How does the answer change if the AI has been personalized (i.e., fine-tuned) on previous outputs produced without AI assistance by the same human? We conducted a preregistered experiment with representative sampling (N = 1802) repeated in four countries (United States, United Kingdom, China, and Singapore). We investigated laypeople's attributions of credit and blame to human users for producing beneficial or harmful outputs with a standard large language model (LLM), a personalized LLM, or no AI assistance (control condition). Participants generally attributed more credit to human users of personalized versus standard LLMs for beneficial outputs, whereas LLM type did not significantly affect blame attributions for harmful outputs, with a partial exception among Chinese participants. In addition, UK participants attributed more blame for using any type of LLM versus no LLM. Practical, ethical, and policy implications of these findings are discussed.
Bilingual Toddlers' Vocabulary Growth Interacts with Existing Knowledge and Cross-Linguistic Similarity.
We explored whether bilingual toddlers make use of semantic and phonological overlap between their languages to learn new words. We analysed cross-sectional and longitudinal CDI data on the words understood and produced by 1.0 to 3.0-year-old bilingual toddlers with English and one additional language. Cognates were more likely to be understood and produced compared to non-cognates. Cognate effects were modulated by whether the toddler knew the translation equivalent in the other language, highlighting that young learners are sensitive to the similarities across their languages. Additionally, exploratory analyses suggest that children with smaller vocabularies rely more on translation equivalents to support the acquisition of difficult words. Children with larger vocabulary sizes exhibited no preference for translation equivalents in comprehension, and a preference for new concepts in production. The rapid acceleration of vocabulary growth in the second year of life may explain this developmental change in translation equivalent preference.
Psychiatric symptoms in COVID-19-positive individuals in the general population: Trajectories of depression, anxiety, and insomnia.
The present study investigates differences in the trajectories of anxiety, depression, and sleep problems among infected versus non-infected case-controlled individuals. Patients who tested positive for COVID-19 were selected from a representative sample in Norway (N > 10,000). In total, 126 of these individuals were infected during the project period, and this group was analyzed at T5 (May 2021). Of these positive cases, those who had completed both PHQ-9 and GAD-7 at all three measurement points were selected for longitudinal analysis using multilevel modeling. There was a significant difference at T5 between those who had tested positive for COVID-19 and matched controls. Anxiety and depression were reduced among those who tested positive, but there were no differences in trajectory when compared to matched controls. Limitations include the use of self-report measures and the assessment of symptoms at a time when strict virus mitigation protocols were in place. The present findings indicate that individuals who test positive for COVID-19 exhibit higher levels of depressive symptoms after restrictions are lifted. However, comparison of anxiety and depression symptom trajectories with matched controls reveals that both groups exhibited stable or slightly decreased symptoms.
Going ‘meta’: a systematic review of metacognition and functional neurological disorder
In functional neurological disorder (FND), there is a fundamental disconnect between an apparently intact nervous system and the individuals’ ability to consistently perform motor actions, perceive sensory signals and/or access effective cognition. Metacognition, the capacity to self-evaluate cognitive performance, appears highly relevant to FND pathophysiology. Poor metacognition is a potential mechanism via which abnormal models of self and the state of the world could arise and persist unchecked. There is therefore a justified enthusiasm that studies of metacognition may give substance to FND’s intangible nature. However, many assume an impairment in metacognition even though experimental studies are still in their infancy. This systematic review provides an analytical checkpoint of the evidence after the first five years of experimental work. We firstly summarize current methods for testing metacognition, prerequisite knowledge that allows readers to independently evaluate the evidence. Using the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines, we then screened the 21 articles on this topic and review the experimental data of the eight studies that specifically tested metacognition in subjects with FND. Questionnaire metrics used to estimate global metacognition and general confidence in FND revealed a mixed picture of low or normal confidence. Of the five studies that used performance-controlled metrics, the gold-standard to estimate local metacognition in FND, four found metacognition to be equivalent to healthy controls and one paper supported impaired metacognition. We consequently try and broaden the debate and discuss alternative headline scenarios. We review how positive studies may offer insight and debate whether null studies could represent false negatives. However, since most studies find equivalent metacognition to controls, we also discuss whether metacognition could be intact and how this could inform mechanistic models of FND and have potential clinical utility. In summary, this review highlights signal of interest within the data, exposes current limitations and flags the many open questions.
Behavioural changes in frontotemporal dementia and their cognitive and neuroanatomical correlates.
Behavioural changes are a central feature of frontotemporal dementia (FTD); they occur in both behavioural-variant (bvFTD) and semantic dementia (SD)/semantic-variant primary progressive aphasia subtypes. In this study, we addressed two current clinical knowledge gaps; (i) are there qualitative or clear distinctions between behavioural profiles in bvFTD and SD, and (ii) what are the precise roles of the prefrontal cortex and anterior temporal lobes in supporting social behaviour? Resolving these conundrums is crucial for improving diagnostic accuracy and for the development of targeted interventions to treat challenging behaviours in FTD. Informant questionnaires to assess behavioural changes included the Cambridge Behavioural Inventory-Revised and two targeted measures of apathy and impulsivity. Participants completed a detailed neuropsychological battery to permit investigation of the relationship between cognitive status (including social-semantic knowledge, general semantic knowledge and executive function) with behaviour change in FTD. To explore changes in regional grey matter volume, a subset of patients had structural MRI. Diagnosis-based group comparisons were supplemented by a transdiagnostic approach which encompassed the spectrum of bvFTD, SD and "mixed" or intermediate cases. Such an approach is sensitive to the systematic graded variation in FTD and allows the neurobiological underpinnings of behaviour change to be explored across an FTD spectrum. We found a wide range of behavioural changes across FTD. Although quantitatively more severe on average in bvFTD, as expected, the item-level analyses found no evidence for qualitative differences in behavioural profiles or "behavioural double dissociations" between bvFTD and SD. Comparisons of self and informant ratings revealed strong discrepancies in the perspective of the caregiver versus patient. Logistic regression revealed that neuropsychological measures had better discriminative accuracy for bvFTD versus SD than caregiver-reported behavioural measures. A principal component analysis of all informant questionnaire domains extracted three components, interpreted as reflecting: (1) apathy, (2) challenging behaviours and (3) activities of daily living. More severe apathy in both FTD subtypes was associated with (a) increased levels of impaired executive function and (b) anterior cingulate cortex atrophy. Questionnaire ratings of impaired behaviour did not correlate with either anterior temporal lobe atrophy or degraded social-semantic knowledge. Together, these findings highlight the presence of a wide range of behavioural changes in both bvFTD and SD, which vary by degree rather than quality. We recommend a transdiagnostic approach for future studies of the neuropsychological and neuroanatomical underpinnings of behavioural deficits in FTD.
Parenting Experiences in the Context of Parental Bipolar Disorder: A Systematic Review and Meta-Synthesis of the Qualitative Literature
Abstract Parental bipolar disorder (BD) is associated with an increased risk of mental health problems in children. Despite the urgent need for clear guidance on how best to support parents with BD, current research lacks a unified analysis of the challenges and needs faced by these parents and their children. This review aims to explore the impact of BD on experiences of parent–child interactions or relationships to inform effective policies and interventions. Following a preregistered PROSPERO protocol, we searched Medline, Embase, PsycINFO, SCOPUS, and CINAHL for qualitative studies on parents with BD and their children (under 19 years) published since 1994. Each study was independently screened and jointly assessed for quality using the Critical Appraisal Skills Program. Our thematic synthesis entailed coding in NVivo, followed by collaborative theme generation on the Miro platform. We reviewed 19 studies, of which 9 studies reported on parents, 8 on children, and 2 on both parents and children. Our analysis generated four themes: (1) "The multifaceted landscape of parenting with BD," which outlines how mood swings affect parenting in diverse ways; (2) "The evolving dynamic of child-parent relationship amidst parental BD," including how children adapt and grow in understanding and responsibility in response to their parent’s BD; (3) "The dual nature of childcare for parents with BD," which explores how childcare impacts parents’ emotions and motivations, bringing both uplifting and challenging effects; (4) "Navigating parental challenges in the context of BD," highlighting the importance of open communication, self-reflection, and timely, unbiased support to mitigate challenges associated with parental BD. This qualitative synthesis focuses specifically on the parenting experiences of families affected by parental BD. It highlights the complex, dynamic impact of BD on parenting behaviors and children’s coping mechanisms, calling for tailored therapeutic interventions that benefit both parents and children. The scope of our study is limited by factors such as a predominance of Western perspectives and an underrepresentation of fathers’ experiences, highlighting the need for more diverse research in this area.
Effects of noise and metabolic cost on cortical task representations.
Cognitive flexibility requires both the encoding of task-relevant and the ignoring of task-irrelevant stimuli. While the neural coding of task-relevant stimuli is increasingly well understood, the mechanisms for ignoring task-irrelevant stimuli remain poorly understood. Here, we study how task performance and biological constraints jointly determine the coding of relevant and irrelevant stimuli in neural circuits. Using mathematical analyses and task-optimized recurrent neural networks, we show that neural circuits can exhibit a range of representational geometries depending on the strength of neural noise and metabolic cost. By comparing these results with recordings from primate prefrontal cortex (PFC) over the course of learning, we show that neural activity in PFC changes in line with a minimal representational strategy. Specifically, our analyses reveal that the suppression of dynamically irrelevant stimuli is achieved by activity-silent, sub-threshold dynamics. Our results provide a normative explanation as to why PFC implements an adaptive, minimal representational strategy.
Investigators are human too: outcome bias and perceptions of individual culpability in patient safety incident investigations.
BACKGROUND: Healthcare patient safety investigations inappropriately focus on individual culpability and the target of recommendations is often on the behaviours of individuals, rather than addressing latent failures of the system. The aim of this study was to explore whether outcome bias might provide some explanation for this. Outcome bias occurs when the ultimate outcome of a past event is given excessive weight, in comparison to other information, when judging the preceding actions or decisions. METHODS: We conducted a survey in which participants were each presented with three incident scenarios, followed by the findings of an investigation. The scenarios remained the same, but the patient outcome was manipulated. Participants were recruited via social media and we examined three groups (general public, healthcare staff and experts) and those with previous incident involvement. Participants were asked about staff responsibility, avoidability, importance of investigating and to select up to five recommendations to prevent recurrence. Summary statistics and multilevel modelling were used to examine the association between patient outcome and the above measures. RESULTS: 212 participants completed the online survey. Worsening patient outcome was associated with increased judgements of staff responsibility for causing the incident as well as greater motivation to investigate. More participants selected punitive recommendations when patient outcome was worse. While avoidability did not appear to be associated with patient outcome, ratings were high suggesting participants always considered incidents to be highly avoidable. Those with patient safety expertise demonstrated these associations but to a lesser extent, when compared with other participants. We discuss important comparisons between the participant groups as well as those with previous incident involvement, as victim or staff member. INTERPRETATION: Outcome bias has a significant impact on judgements following incidents and investigations and may contribute to the continued focus on individual culpability and individual focused recommendations observed following investigations.
Early-stage randomised controlled trial of therapist-supported online cognitive therapy for post-traumatic stress disorder in young people.
BACKGROUND: Effective face-to-face treatments for Post-Traumatic Stress Disorder (PTSD) are available, but most young people with PTSD do not receive effective treatment. Therapist-supported online Cognitive Therapy has the potential to improve accessibility of effective treatment. This early-stage trial gathered data on the feasibility, acceptability, and initial signal of clinical efficacy of a novel online Cognitive Therapy program for young people with PTSD. METHODS: A two-arm, parallel-groups, single-blind, early-stage feasibility RCT compared online Cognitive Therapy to a waitlList condition. Participants were N = 31 adolescents (12-17 years-old) with a diagnosis of PTSD, randomised in a 1:1 ratio using minimisation. Thresholds for progression to a larger trial were set a priori for recruitment rate, data completeness, and the initial signal of clinical efficacy. The primary clinical outcome was PTSD diagnosis at 16 weeks post-randomisation. Secondary clinical outcomes were continuous measures of PTSD, depression, and anxiety at 16 weeks; and at 38 weeks in the online Cognitive Therapy arm. RESULTS: All pre-determined feasibility thresholds for progression to a larger trial were met. We recruited to target at a rate of 1-2 participants/month. No patient dropped out of therapy; 94% of all participants were retained at 16 weeks. At 16-weeks, the intention-to-treat (ITT) effect adjusted odds ratio was 0.20 (95% CI, 0.02, 1.42), indicating that the odds of meeting PTSD caseness after online therapy were 80% lower than after the waitlist (10/16 participants met PTSD caseness after therapy compared to 11/13 after WL). Effect-size estimates for all secondary clinical outcomes were large-moderate; improvements were sustained 38 weeks after online Cognitive Therapy. CONCLUSIONS: Therapist-supported online Cognitive Therapy for PTSD is acceptable to young people and has potential for meaningful and sustained clinical effects. A larger trial appears feasible to deliver. Further work is needed to refine the intervention and its delivery and to evaluate it in a larger confirmatory trial.