Medical domain knowledge in domain-agnostic generative AI

The text-guided diffusion model GLIDE (Guided Language to Image Diffusion for Generation and Editing) is the state of the art in text-to-image generative artificial intelligence (AI). GLIDE has rich representations, but medical applications of this model have not been systematically explored. If GLIDE had useful medical knowledge, it could be used for medical image analysis tasks, a domain in which AI systems are still highly engineered towards a single use-case. Here we show that the publicly available GLIDE model has reasonably strong representations of key topics in cancer research and oncology, in particular the general style of histopathology images and multiple facets of diseases, pathological processes and laboratory assays. However, GLIDE seems to lack useful representations of the style and content of radiology data. Our findings demonstrate that domain-agnostic generative AI models can learn relevant medical concepts without explicit training. Thus, GLIDE and similar models might be useful for medical image processing tasks in the future - particularly with additional domain-specific fine-tuning.

The Cambridge Handbook of Lifespan Development of Creativity

This handbook focuses on the development and nurturance of creativity across the lifespan, from early childhood to adolescence, adulthood, and later life. It answers the question: how can we help individuals turn their creative potential into achievement? Each chapter examines various contexts in which creativity exists, including school, workplace, community spaces, and family life. It covers various modalities for fostering creativity such as play, storytelling, explicit training procedures, shifting of attitudes about creative capacity, and many others. The authors review research findings across disciplines, encompassing the work of psychologists, educators, neuroscientists, and creators themselves, to describe the best practices for fostering creativity at each stage of development.

Explicit versus non-explicit prosodic training in the learning of Spanish L2 stress contrasts by French listeners

Different methods to acquire a language can contribute differently to learning success. In the present study we tested the success of L2 stress contrasts acquisition, when ab initio learners were taught or not about the theoretic nature of L2 stress contrasts. In two 4-hour perceptual training methods, French-speaking listeners received either (a) explicit instructions about Spanish stress patterns and perception activities commonly used in L2 pronunciation courses or (b) no explicit instructions and a unique perception activity, a shape/word matching task. Results showed that French-speaking listeners improved their ability to identify and discriminate stress contrasts in Spanish after training. However, there was no significant difference between explicit and non-explicit training nor was there an effect on stress processing under different phonetic variability conditions. This suggests that in L2 stress acquisition, non-explicit training may benefit ab initio learners as much as explicit instruction and activities used in L2 pronunciation courses.

Biased Data, Biased AI: Deep Networks Predict the Acquisition Site of TCGA Images

Deep learning models applied to healthcare applications including digital pathology have been increasing their scope and importance in recent years. Many of these models have been trained on The Cancer Genome Atlas (TCGA) atlas of digital images, or use it as a validation source. This study shows that there are tissue source site (tss) specific patterns of TCGA images that could be used to identify contributing institutions without any explicit training. Furthermore, it was observed that a model trained for cancer subtype classification has discovered such tss specific patterns within digital slides to classify cancer types. Digital scanner configuration and noise, tissue stain variation and artifacts, and source site patient demographics are among factors that likely account for the observed bias. Therefore, researchers should be cautious of such bias when using histopathology datasets for developing and training deep networks.

Qualitative Investigation into Therapists’ Experiences of Online Therapy: Implications for Working Clients

Online therapy has increasingly been utilised during the COVID-19 pandemic by many, including working populations. However, few qualitative studies have explored how online therapy is experienced in practice and discussed its implications for those working clients. Semi-structured interviews attended by nine integrative psychotherapists practising in California, the United States, were conducted. Thematic analysis of the transcripts identified three themes: (i) ‘Positive experiences of online therapy’, (ii) ‘Challenges experienced by therapists and clients in online therapy’, and (iii) ‘Preparation and training for online therapy’. Online therapy was assessed as being helpful, particularly in terms of mitigating against previous geographical and temporal barriers to uptake. However, due to technological disruptions and potential blurring of professional boundaries, online therapy may detract from the emotional salience of therapy, negatively impacting the therapeutic relationship and containment. Considering these positive experiences, participants expected that the demand for online therapy would continue to increase. Particularly in the occupational context, online therapy can offer interventions without fostering shame regarding mental health. The findings provide preliminary qualitative evidence that online therapy can be a useful adjunct to traditional forms of face-to-face therapy. However, therapists require more explicit training in implementing online therapy. Results are discussed in particular regarding the utility of this therapy for working clients.

Phonological Awareness of L1 Systemic Segmental Contrasts among Advanced ESL Speakers with Varied L1 Backgrounds

The paper explores the phonological awareness of L1 among advanced adult speakers of EFL in the context of L2 pronunciation training. The subjects are students of English with Polish, Spanish, Turkish and Russian L1 background. All subjects have participated in intensive English pronunciation instruction as part of their degree training, in the English Department at the Pedagogical University in Kraków. Two aspects are tar- geted for examination: perception of sound contrasts and awareness of contextual variants in L1, mostly those pertaining to the consonantal and vocalic inventories, all related to their L2 (English) production goals. The material is based on longitudinal examination of course test results over the span of 3 years. The analysis reveals low sound discrimination skills in the subjects’ L1, largely based on letter-to-sound correspondences and inability to see beyond print. Through explicit training in their L2 they become more sensitive to the inventory and the details of their L1 sound system, the awareness they can use to the advantage when targeting L2 sound production.

Explicit Training to Improve Affective Prosody Recognition in Adults with Acute Right Hemisphere Stroke

Difficulty recognizing affective prosody (receptive aprosodia) can occur following right hemisphere damage (RHD). Not all individuals spontaneously recover their ability to recognize affective prosody, warranting behavioral intervention. However, there is a dearth of evidence-based receptive aprosodia treatment research in this clinical population. The purpose of the current study was to investigate an explicit training protocol targeting affective prosody recognition in adults with RHD and receptive aprosodia. Eighteen adults with receptive aprosodia due to acute RHD completed affective prosody recognition before and after a short training session that targeted proposed underlying perceptual and conceptual processes. Behavioral impairment and lesion characteristics were investigated as possible influences on training effectiveness. Affective prosody recognition improved following training, and recognition accuracy was higher for pseudo- vs. real-word sentences. Perceptual deficits were associated with the most posterior infarcts, conceptual deficits were associated with frontal infarcts, and a combination of perceptual-conceptual deficits were related to temporoparietal and subcortical infarcts. Several right hemisphere ventral stream regions and pathways along with frontal and parietal hypoperfusion predicted training effectiveness. Explicit acoustic-prosodic-emotion training improves affective prosody recognition, but it may not be appropriate for everyone. Factors such as linguistic context and lesion location should be considered when planning prosody training.

Timing storytime to maximise children's ability to retain new vocabulary

Shared storybook reading is a key aid to vocabulary acquisition in childhood. However, word learning research has tended to use unnaturalistic (explicit) training regimes. Utilising a storybook paradigm, we examine whether children (particularly those with weaker vocabulary) are more likely to retain new words if they learn them closer to sleep. Parents read their child (5-7ys, n=237) an alien adventure story that contained 12 novel words with illustrations, at one of two training times: at bedtime or 3-5hrs before bedtime. Using on-line tasks, parents tested their child’s ability to recall the new words (production) and associate them with pictures (comprehension), immediately after hearing the story and again the following morning. As hypothesized, we replicated two findings: children showed overnight improvements in their ability to produce and comprehend new words when tested again the next day, and children with better existing vocabulary knowledge showed larger overnight gains in new word comprehension. Counter to expectations, overnight gains in comprehension were larger if the story was read 3-5 hours before, than at, bedtime. These ecologically valid findings are consistent with theories that characterise word learning as a prolonged process supported by mechanisms such as consolidation and retrieval practice, with existing vocabulary knowledge acting as an important source of variability in retention. The findings provide preliminary evidence that encountering new words in stories later in the day (but not too close to sleep) may help to harness vocabulary growth and may be more beneficial than leaving shared storybook reading just for bedtime.

Autonomy Orientations of English Language Teachers in Turkey

The present study aimed to explore autonomy orientations of English language teachers' work at an intensive language school in Turkey. The autonomy orientations of the teachers were analysed through self-determination theory. The study also investigated if these orientations vary according to the teachers’ gender, years of experience, and department of graduation. The investigation was carried out through an autonomy orientations questionnaire. Data were collected from 111 language teachers, 11 of whose opinions were utilized for further analysis. The results showed that teachers had moderate autonomy supportive orientation, which was reported to be risky. Moreover, gender was the only variable that had a significant effect on the autonomy orientations. The teachers mainly blamed the education system for restricting their adoption of the autonomous orientations. The results imply the necessity of explicit training on how teachers can be encouraged to have more autonomy supporting orientations.

The Communication Benefits of Participation in Camp Dream. Speak. Live.: An Extension and Replication

School-based guidelines often require that treatment focuses on minimizing or eliminating stuttered speech. The purpose of this study was to examine the benefits of explicit training in communication competencies to children who stutter without targeting stuttered speech. Thirty-seven children (ages 4–16) completed Camp Dream. Speak. Live., an intensive group treatment program which targets the psychosocial needs and communication of children who stutter. Outcome measures included the Overall Assessment of the Speaker's Experience of Stuttering (OASES), the Communication Attitude Test for Preschool and Kindergarten Children Who Stutter (KiddyCAT), and the Patient Reported Measurement Information System (PROMIS) Pediatric Peer Relationships Short Form (PROMIS Peer Relationships) and Parent Proxy Peer Relationships Short Form (PROMIS Parent Proxy). Pre- and posttreatment public presentations were rated on nine core verbal and nonverbal communication competencies by a neutral observer. Similar to previous studies, participants demonstrated significant improvements in communication attitudes (OASES) and perceived ability to establish peer relationships (PROMIS Peer Relationships), particularly school-aged participants (ages 7–16). Participants also demonstrated significant improvement in eight of the nine communication competencies. Findings suggest that, in addition to the psychosocial gains of programs such as Camp Dream. Speak. Live., children who stutter benefit from explicit training in communication skills, and these gains are not dependent on the presence of stuttered speech.