Behavior Research Methods

Purpose: To evaluate the feasibility, validity, and scalability of PLFest, an open-source, Unity-based, cross-platform application designed for standardized, multi-site visual and cognitive assessment and training. Methods: Two hundred sixty participants (mean age = 23 years) were recruited across four university sites in the United States. Participants completed a battery of five visual assessments administered through PLFest, including visual acuity, contrast sensitivity, spatial frequency cutoff, contrast sensitivity at spatial-frequency cutoff, and visual search. Five cognitive assessments measuring visuospatial working memory, verbal working memory, fluid reasoning, inhibitory control, and selective attention were also administered. Descriptive statistics and performance distributions were examined and compared with normative data. Results: Visual acuity and contrast sensitivity measures closely matched previously reported normative values obtained using established clinical and psychophysical methods. Spatial frequency cutoff and visual search tasks produced stable threshold estimates while showing substantial inter-individual variability. Performance across all cognitive assessments was consistent with published validation studies of the corresponding tasks. Across the full battery, adaptive procedures demonstrated reliable convergence and generated well-distributed performance measures without evidence of substantial floor or ceiling effects. Importantly, these findings were observed across four geographically distributed testing sites using standardized consumer-grade tablet hardware. Conclusions: PLFest provides reliable and scalable assessment of visual and cognitive function using portable consumer devices. The platform supports standardized data collection across distributed research settings while maintaining performance characteristics consistent with established laboratory and clinical benchmarks. These findings support the use of PLFest as a reliable framework for large-scale studies of vision and cognition. Translational Relevance: By reducing dependence on specialized laboratory infrastructure and trained personnel, PLFest may facilitate broader access to visual and cognitive assessment, enabling large-scale research, screening, and future rehabilitation applications.

Joint music making offers an ecologically powerful framework for investigating human social interaction and synchronization. Yet, experimental paradigms often rely on traditional instruments that limit accessibility, reproducibility, and experimental control. In parallel, the use of music for therapy and rehabilitation is expanding, motivating the development of digital musical instruments that can serve research, educational, and clinical purposes. Here, we introduce the e-Music Box Roma (eMB Roma), an open, reproducible digital musical instrument designed to study music making behavior regardless of musical training. The eMB Roma plays preregistered music with tempo controlled by hand rotary movements. Building on the original e-Music Box (Novembre et al., 2015), the eMB Roma retains its intuitive rotary hand control while introducing major innovations: a fully open and 3D-printable design, modular hardware with integrated slider and button controls, polyphonic output with multiple simultaneous instruments, and MIDI compatibility. Additionally, a dedicated graphical user interface allows real-time monitoring, experiment control, device synchronization (like neuroimaging or motion capture devices), and both solo and joint music-making paradigms. The eMB Roma provides a flexible and accessible platform for research contexts, allowing experimental control, reproducibility, and future extensions. Its open design and modularity make it suitable not only for research but also for therapeutic, rehabilitation, and educational applications, where it can support personalized interventions and quantitative assessment of motor performance.

Cognitive impairment is a prevalent symptom extending from physiological ageing to disease. It commonly manifests itself in initial memory problems, progressing and co-occurring in more severe conditions such as Mild Cognitive Impairment, Alzheimer's Disease and Major Depressive Disorder. However, current non-invasive screening assessments either lack biological information or are invasive and restricted to specialized centers with complex and cost-intensive set-ups. Here, we conducted an initial validation of mobile pupillometry with Virtual Reality (VR) under experimental conditions as a digital biomarker for cognitive impairment by testing required biomarker-specific properties. For this purpose, we first assessed its construct validity by testing healthy participants (n=43) on an n-back task in VR while pupil size was measured. Mixed effects models revealed that similar to lab-based eye-tracking systems, pupil size increased in a sensible and distinguishable fashion as a function of working memory load. Second, to test the signal's reliability, the same participants were tested on the identical set-up two to three months after their first visit. We observed that the pupil response profile was highly stable over this period. Third, for its clinical validity, we examined patients (n=89) from three different cohorts with varying degrees of cognitive impairment and compared them to healthy control participants (n=81). Mixed-effects models indicated that pupil size was reduced as a function of cognitive impairment levels at higher cognitive load and that this effect was stronger pronounced with increasing age. In conclusion, we provide initial evidence for mobile pupillometry being a sensitive, reliable and clinically valid digital biomarker for cognitive functioning and impairment, which offers desirable properties due to its quick, automatized and location-independent set-up. Keywords: digital biomarker, mobile pupillometry, Virtual Reality, cognition, , Major Depressive Disorder, Mild Cognitive Impairment, Alzheimer's Disease

1. Video-tracking programs have now become an essential tool for researchers measuring animal behavior across biological fields. The panel of available programs is growing rapidly, providing researchers with numerous specific tools that will match their precise needs. However, their proliferation may complicate post-tracking data processing, and some programs do not even provide tools for correcting tracking errors or analysing tracking data. In the case of commercial software, the loss of access to a program due to budget limitations or researchers' mobility from one institution to another could prevent them from accessing and visualizing their tracking data. 2. There is therefore a growing need for an accessible and flexible tool to handle post-tracking processes such as the correction and analysis of tracking data obtained across different video-tracking programs. 3. We present here the latest update of the video tracking and analysis program AnimalTA. With this new release, we propose to solve the above-mentioned problems by providing the scientific community with a program that will allow for data importation from other video-tracking programs. Like in its previous versions, AnimalTA remains a free, open-source, and highly user-friendly program, ensuring that it will always be accessible without restriction. Now, with this new importation option, users who performed their tracking with other programs can benefit from AnimalTA's complete toolset of data visualization, correction, and analysis. 4. Finally, this article gives an overview of the other main improvements associated with this new release. The program is now faster in both video importation and tracking, proposes an amplified toolset for data visualisation and correction, and features new options for data analysis.

Visual search is a critical component of many professions such as military operations, baggage screening, and radiology. Aided Target Recognition (AiTR) systems are designed to highlight potential threats across the operator visual field in real-time, directing attention and improving accuracy. However, these systems may impact search and, consequently, situational awareness by diverting attentional resources from non-highlighted, yet relevant, locations. Previous work suggests that scene gist is extracted within the first 250 ms of scene onset (Vo & Henderson, 2010). As such, this study examined whether a 250 ms AiTR onset delay could encourage a more even distribution of attention. Participants searched synthetically generated scenes and classified each person in the scene as armed or unarmed. Depending on their condition, participants either saw the scenes unaugmented (No AiTR condition), with AiTR highlights consisting of red bounding boxes around armed people and yellow boxes around unarmed (AiTR condition), or with AiTR highlights presented 250 ms post scene onset (Delayed AiTR condition). A surprise memory test of background objects presented in the search scenes was administered to all participants upon completion of the search task. As predicted and preregistered, results showed less overt attentional deployment to background information (anything other than the people themselves) in the AiTR condition compared to No AiTR , however, decreased overt attentional deployment was not seen in the Delayed AiTR group. A similar pattern was observed in the memory data (with the AiTR condition having a lower score than the No AiTR condition and the Delayed AiTR condition), this difference was not significant.

Background: Motor disturbances are common in neurologic and neurodegenerative syndromes. A standard motor speed and dexterity measure is the finger tapping test (FTT). The FTT has traditionally been administered in clinic using a mechanical FTT, limiting accessibility and early motor change quantification. This study assessed the validity of a smartphone app-based FTT, which may expand access and enable more frequent testing. Methods: The cohort was diagnostically diverse, including participants with frontotemporal dementia (FTD), progressive supranuclear palsy (PSP), corticobasal syndrome, primary progressive aphasia, multiple sclerosis, and clinically unimpaired controls. Participants completed a 20-second ALLFTD Mobile App (mApp)-FTT with each hand. Tapping speed metrics were extracted. Participants completed the gold-standard mechanical FTT, a neurologist-administered finger tapping exam, the PSP Rating Scale (PSPRS) and the Unified Parkinson`s Disease Rating Scale (UPDRS). Correlations assessed mApp-FTT and mechanical FTT relationships; regressions evaluated associations with neurologist-rated finger tapping impairment, PSPRS and UPDRS, adjusting for age and sex. Results: The mApp-FTT showed moderate-to-strong correlations with the mechanical FTT (dominant: r=0.63, p<0.001; non-dominant: r=0.55, p<0.001). Taps per second were associated with PSPRS motor severity (dominant hand: std. {beta}=-0.59, 95% CI [-0.91, -0.27], p<0.001) and the UPDRS (dominant hand: std. {beta}=-0.41, 95% CI [-0.82, 0.00], p=0.049). Flight time was modestly associated with neurologist-rated finger tapping impairment (dominant hand: std. {beta}=0.15, 95% CI [0.00, 0.29], p=0.044). Conclusion: These findings support mApp-FTT validity as a measure of motor function across neurodegenerative conditions. Validation in longitudinal and unsupervised remote settings is warranted to understand scalability and evaluate change over time.

Speech comprehension in noisy environments often requires cognitive effort, but listeners may disengage when comprehension becomes impossible. Eye movements have recently emerged as a promising new measure of listening effort, but it remains unclear whether eye movements are sensitive to the full effort profile across easy, difficult, and impossible speech comprehension. Across four experiments, participants listened to sentences at easy, difficult, and impossible levels of multi-talker background babble while pupil size and eye movements were recorded. Pupil size generally followed the expected inverted u-shaped effort profile: low for easy speech, maximal for difficult but still intelligible speech and lower again for impossible speech, although this pattern partly reflected sustained, condition-specific differences and not only sentence-evoked responses. Gaze dispersion - measuring the spread of eye movements - decreased with high temporal selectivity during difficult relative to easy and impossible speech, indicating reduced eye movements during active, effortful listening. However, gaze dispersion was also lower, but less temporally selective, during impossible compared to easy listening, especially in non-baseline-corrected analyses, suggesting that reduced eye movements do not index listening effort uniquely. Instead, eye movements appear to reflect both attentional engagement during difficult listening and disengagement or inward attention when meaningful listening is no longer possible. These findings indicate that pupil size and eye movements provide complementary indices of listening-related cognition, and highlight the integration of listening, cognition, and motor systems.

This study examined the extent to which goodness of pronunciation (GoP) scores and phonological posterior probabilities capture perceptual ratings of speech severity in individuals with motor speech disorders (MSD). Speech recordings of the word catastrophe were obtained from 489 participants, including 333 neurologically typical controls and 156 individuals with MSD. GoP scores were derived using traditional acoustic features and self-supervised speech representations, including WavLM and XLS-R, across multiple modeling approaches, while phonological posterior probabilities were extracted using Phonet. Model performance was evaluated using Kendall's rank correlations, regression, and receiver operating characteristic analyses against speech-language pathologists' perceptual ratings of sound distortion and intelligibility. Both GoP and phonological posterior probabilities were significantly associated with perceptual ratings. Self-supervised speech representations substantially outperformed traditional acoustic features, with WavLM-based GoP using k-nearest neighbors achieving the strongest performance. Across correlation, regression, and classification analyses, GoP consistently outperformed phonological posterior probabilities for both sound distortion and intelligibility. Age and gender had minimal influence on model-derived measures or their relationships with perceptual ratings. These findings demonstrate the value of self-supervised GoP as an objective measure of speech impairment while highlighting the complementary role of phonological posterior probabilities in characterizing articulatory aspects of motor speech disorders.

A major challenge in behavioral neuroscience is the lack of a unified software framework capable of implementing diverse paradigms across species and experimental setups. Researchers currently face a trade-off: they must either spend significant time developing custom, siloed solutions that hinder reproducibility, or incur substantial costs purchasing inflexible, closed systems. Here, we present Neurokraken, an open-source, Python-native platform designed to overcome these limitations. Neurokraken allows writing experiment progression entirely in standard python, while its core architecture automatically sets up a microcontroller for the connected hardware components and enables python side access with millisecond-precision timing and automatic logging. The system prioritizes ease of use and flexibility, enabling advanced series of events and conditions, the usage of python ecosystem code and packages within experiments, and the addition of any arduino-compatible electronic devices for custom experiments. As a result, users can easily create interactive virtual and real environments to engage, monitor, and record subjects. We present Neurokraken's versatility across a wide range of paradigms, for human and non-human primate psychophysics, and complex rodent behavior in both head-fixed and freely moving paradigms. Its modular design allows for rapid hardware reconfiguration, while a fully customizable user interface enables real-time monitoring and interactive experimental control without compromising timing precision. By uniting laboratory-grade precision with an accessible and flexible open-source philosophy, Neurokraken provides a single, powerful solution to design and execute next-generation behavioral experiments. We hope Neurokraken helps accelerate research, improve reproducibility throughout the neuroscience community, and make advanced behavioral experimentation more accessible through its substantial cost-efficiency.

ObjectiveThe Lab Streaming Layer (LSL) protocol, widely adopted for synchronized multimodal biosignal recording in neuroscience research, transmits all data in plaintext, exposing sensitive neural and physiological recordings to interception and tampering and creating regulatory liability for clinical and commercial deployments across most major international jurisdictions. We present Secure LSL, the unified encryption architecture for the protocol: a novel security layer that authenticates devices and encrypts biosignal streams through transparent, drop-in modifications to the core library, requiring no application changes and no recompilation for dynamically linked clients. ApproachWe implement encryption at the liblsl core library level using a shared keypair authorization model with ChaCha20-Poly1305 authenticated encryption. All authorized devices share a common Ed25519 keypair, and public key verification during connection establishment ensures only authorized devices communicate. The architecture enforces network-wide security consensus, requiring all connected devices to operate in either secure or insecure mode, eliminating vulnerable mixed environments, and operates transparently with zero code changes to existing applications. Main ResultsThe architecture preserves application programming interface (API) transparency, so existing applications need no code changes (legacy devices must update to connect to secured outlets). Across five hardware platforms spanning x86 desktop, Apple Silicon laptop, embedded ARM single-board, and Xtensa microcontroller targets, encryption adds sub-millisecond latency in all desktop and embedded ARM configurations, with overhead in the single-digit percent range (approximately 4 to 9%, the lowest values within measurement noise of zero) for typical 64-channel, 1000-Hz configurations. A clean-room ESP32 implementation extends transparent encryption to dual-core microcontrollers with no measurable push-path overhead and approximately 2 kB additional static random-access memory (SRAM) consumption, enabling secured wearable and ambulatory biosensor deployments. SignificanceBy implementing security within the protocol core rather than requiring application-level changes, we transform LSL from a research-only protocol to a security-capable platform for clinical settings, multi-institution collaborations, and commercial products, while preserving its zero-configuration philosophy.

Digital technologies can support meaningful social interaction by providing personally relevant prompts for memory, communication, and shared reflection. In later life, mobile phone photography may offer an accessible medium through which older adults and caregivers construct stories, express meaning, and participate in relational engagement. However, limited psychophysiological evidence is available on how digital photo supported storytelling engages cognitive and social processes in older adult caregiver dyads. This study examined alpha band EEG dynamics during digital photo elicited storytelling in two museum settings. Thirty two older adult caregiver dyads completed cognitive and psychological screening and participated in a museum-based storytelling protocol. During the museum visit, participants used mobile-phone photography to capture personally meaningful objects, scenes, or exhibition spaces. Each participant then selected one photograph as a digital prompt for a structured but naturalistic storytelling interaction. EEG was recorded during eyes closed resting, eyes open resting, storytelling, and listening conditions. Relative alpha power was analyzed using a predefined 10 electrode sensor level set. Task related alpha modulation was examined relative to eyes open resting. Associations between Cz alpha power and MoCA scores were tested, and dyad level alpha band inter brain similarity was explored using spatial alpha power patterns with within site shuffled dyad surrogate comparisons. Alpha power was higher during eyes closed resting and lower during storytelling and listening relative to eyes-open resting, indicating task-related alpha modulation during digital photo supported narrative interaction. Associations between MoCA scores and Cz alpha power were weak, condition-specific, and did not survive false discovery rate correction. During storytelling, dyad level alpha-band inter brain similarity was modestly higher than within site shuffled dyad estimates, but this effect did not remain significant after correction across conditions. These findings suggest that digital photo elicited storytelling can provide a meaningful medium for studying cognitive and social engagement in older adult caregiver dyads. Alpha band EEG activity was sensitive to storytelling and listening, although cognition related and dyadic similarity effects were modest. The study contributes to research on technology supported human behavior by showing how digital image prompts can structure naturalistic social interaction while enabling psychophysiological measurement in real-world contexts.

Flower-visiting insect populations are declining since the 1990s, especially because of the decrease of floral resources in agricultural settings. Mass flowering crops can help increase resource availability, and plant breeding can be directed towards selecting varieties attracting more flower-visiting insects. This requires the implementation of an automated high-throughput phenotyping tool for assessing the attractiveness of plant genotypes to flower-visiting insects. In this study, (i) we present a procedure to take standardized images of sunflower heads with camera traps continuously at day and night in the field; (ii) we trained two versions of a deep learning model, named PolliCrop, to automatically detect and identify three classes of the main insects visiting sunflower on these images (non-Bombus bees, bumble bees, lepidopterans); (iii) we assessed and validated the ability of PolliCrop to correctly predict the true visitation frequencies of the insect classes on three sunflower genotypes; (iv) we presented two statistical approaches to compare the insect visitation frequencies between plant genotypes, one including weather variables, and the other one without. One PolliCrop version yielded satisfying performance to correctly detect the three insect classes. In particular, it correctly predicted the insect visitation frequencies on two sunflower genotypes in a range of {+/-}10%. The other PolliCrop version can be useful in certain contexts of images and objectives. PolliCrop can be extended in the future to other crop species by training PolliCrop on new images captured in these crops. The field experimental design to set up for comparing the attractiveness between genotypes is also discussed.

Virtual reality (VR) interaction remains largely dependent on explicit motor input, limiting seamless and adaptive interaction. This study investigated whether electroencephalography (EEG)-based passive brain-computer interfaces (BCIs), combined with eye gaze, can decode interaction intent directly from its underlying neurophysiological correlates during dynamic VR gameplay. We operationalized interaction intent as comprising two components: affordance-related evaluation, indicating whether an attended object affords interaction, and approach-avoidance evaluation, indicating the directional tendency of interaction toward desirable or undesirable outcomes. Twenty-three participants completed a VR game with two calibration sessions and one online BCI session. Offline analyses showed above-chance decoding of the binary approach-avoidance decision classification across all actionable trials, with a grand-average accuracy of 66.28% across participants. This decoding transferred to online closed-loop gameplay, where grand-average accuracy remained above chance at 69.64%. Category-level analyses further revealed substantial variability in classification separability. For approach-avoidance-related classifications, accuracy reached 80.84% for the most distinct pairing between clearly valenced reward and punishment categories, but dropped to near chance at 59.03% for the more context-dependent pairing with ambiguous motivational valence. Affordance-related classifications between non-actionable and actionable item categories were consistently high, ranging from 77.76% to 83.50%. User Experience questionnaire results showed that, despite limitations leading to perceived loss of control and reduced ease of use, participants found the BCI-based interaction paradigm itself more fun than the controller baseline. To our knowledge, this is the first demonstration of real-time EEG decoding of interaction intent during dynamic VR gameplay, contributing toward intuitive user-adapted interfaces driven by physiological signals in immersive environments.

In non-verbal communication, observers infer emotions from visible facial movements, yet emotional experiences are described in internal bodily terms (e.g., "my heart skipped a beat"). This contrast highlights a tension between external sensory cues and internal signals. In this context, we examined an overlooked gap in affective science: what makes an emotional portrayal believable, and do believability judgments reflect only what observers can see or also the portraying person's internal cardiac dynamics? To test this, we created 311 scenario-driven acting clips designed to avoid prototypical posed displays. For each clip, we quantified facial movement magnitude from the video, recorded ECG during preparation and enactment, and collected actors' self-reports. Online participants (N = 371) viewed these clips and provided emotion recognition responses and continuous ratings of believability, valence, or arousal. The results show that believability decreased as movement magnitude increased, with a non-linear relationship indicating a stronger penalty as motion increased. Valence further shaped this pattern, with increasing movement reducing believability more strongly for portrayals with negative valence. This effect persisted after accounting for intended emotion, perceived arousal, and emotion recognizability. Cardiac dynamics varied during performance, and actors' higher heart rate variability was associated with higher believability for positively valenced portrayals. Together, these findings show that believability is driven by visible movement cues interpreted in relation to valence, with actors' cardiac dynamics showing selective alignment with believability. These results identify core components of believable emotional expressions and provide a basis for studying such judgments in everyday social interaction.

Introduction: Narrative recall is widely used to detect cognitive impairment, but dominant instruments carry proprietary restrictions. The Craft Story 21 (CS), the non-proprietary NACC UDS4 standard, is not available standalone. Here, we validate the freely available Puppy Escape (PE). Methods: 346 participants (153 cognitively normal, 106 subjective cognitive impairment, 87 mild cognitive impairment) completed PE and CS. Analyses evaluated convergent and criterion validity, MCI-vs-control discrimination, and incremental validity. Results: PE and CS converged (r=.43-.47) and were equivalent on 10/12 neuropsychological measures. PE Delayed discriminated MCI from controls (d=1.03; ROC-AUC equal to CS, DeLong p=.510) and added variance beyond CS (R2=+.054, p<.001). Automated subscores revealed MCI deficits in location, action, and name content. PE-18 short form retained discrimination (d=1.02) with 18 items. Discussion: PE matched CS across all validation domains and captured complementary diagnostic information. PE and PE-18 are available via online registration explicitly permitting industry-sponsored research and fee-for-service clinical use.

Current debates regarding the relative contribution of saliency versus semantics to gaze control often rely on comparing the predictive power of saliency and meaning maps. We argue that such indirect, global approaches are fundamentally limited because fixations arise from heterogeneous, local causes that are conflated in whole-scene comparisons. To substantiate this claim, we used a direct method where participants explicitly identified the reasons for fixation at specific clusters of high fixation density, distinguishing between low-level saliency and various semantic categories, as well as the most important one. The obtained judgments revealed that multiple factors contribute simultaneously to gaze control. Although their influence varied across fixation clusters, semantics generally dominated saliency. Notably, abstract semantic categories, particularly "unknown/unusual," proved important, highlighting the role of prior knowledge and novelty besides personal relevance in guiding attention. To interpret these findings in the context of existing models, we propose a framework distinguishing between processes highlighting interesting locations in the image from a sampling strategy translating this information into scanpaths. Within this framework, classic saliency and meaning maps are viewed as restricted inputs to the strategy, whereas deep learning-based models (e.g., DeepGaze IIE) are more general and may also implicitly encode aspects of the strategy itself. Consistent with this, we found that the predictive performance of DeepGaze IIE varied less significantly with the specific reasons for fixation than that of classic saliency and meaning map approaches.

Neurophysiology datasets are becoming increasingly complex, combining behavioral measurements with high-dimensional neuronal activity recordings coming from optical and/or electrophysiological measurements. The Neurodata Without Borders (NWB) standard has emerged in the community as the format of record. While standardized and widely used preprocessing tools generating NWB files have been developed, extensible frameworks for scientific analysis downstream of the NWB ecosystem are still under-represented. We present CICADA, a Python framework dedicated to analysis of neurophysiological data in the standardized NWB format. The toolbox is built as three hierarchically-organized packages: cicada-nwb (NWB access layer), cicada-analysis (plugin-based analysis engine and tool library), and cicada-gui (PyQt5 desktop application at the head of the pipeline). Beyond this architectural separation, CICADA is built around a central design principle: supporting a continuum from turnkey use to full modularity. Researchers can use the complete GUI-driven cicada-gui workflow without writing code, programmatically use existing analysis plugins from cicada-analysis, contribute to new analysis plugins, reuse utilities from cicada-tools, or build entirely custom pipelines on top of the cicada-nwb access layer alone. The same analysis plugin runs identically in interactive GUI and parameter-configured headless modes, enabling reproducible multi-session, multi-animal group analyses. We illustrate the versatility of CICADA with example analyses of behavioral, calcium imaging (two-photon and widefield) and extracellular electrophysiology datasets from rodent laboratories. CICADA is open source, actively maintained, and designed so that any laboratory can contribute at any level of the stack without modifying the core framework.

Background: Documentation burden significantly impacts nursing workload and well-being, with nurses spending an estimated 20-40% of their time on documentation. Ambient AI technologies offer potential to reduce documentation time by mapping real-time nurse-patient conversations to structured EHR data entries with human-in-the-loop verification. Methods: This protocol describes a pragmatic, EHR-embedded randomized controlled trial evaluating the effectiveness of an Ambient AI tool in reducing nursing documentation time across three inpatient medical/surgical units. The study employs a closed-cohort, stepped-wedge, unit-randomized design, integrating the intervention into routine clinical workflows. The primary outcome is documentation time per shift hour, derived from EHR audit logs. Secondary outcomes include documentation burden, professional well-being, and perceived usability. Results: The trial is being implemented within a shared governance model that integrates executive oversight, operational feasibility, and research rigor. Multidisciplinary workgroups coordinate technical integration, user experience, and analytics, ensuring alignment between operational priorities and pragmatic trial objectives. Early implementation has highlighted the importance of adapting training and analytic strategies to address differential intervention exposure, as well as the need for rapid operational responses to late-emerging technical issues. Discussion: This protocol demonstrates the feasibility of embedding a randomized pragmatic trial within a health system-led operational deployment of Ambient AI for inpatient nursing documentation. The approach highlights the necessity of adapting existing outpatient provider-focused AI implementation strategies for inpatient nursing, emphasizing the unique nature of different nursing care environments. Recruitment challenges and the integration of research with operational workflows are discussed as key considerations for future pragmatic AI trials in nursing. Keywords: Artificial Intelligence; Ambient AI; Nursing Documentation; Documentation Burden; Large Language Models; Speech Recognition Software; Stepped-Wedge Design ClinicalTrials.gov Identifier NCT07456241V4: 2026-05-27 https://clinicaltrials.gov/study/NCT07456241

Auditory brainstem responses (ABRs) are a core part of objective functional evaluations of hearing sensitivity and subcortical auditory transmission. Manual assessments of ABR waveforms are still a primary means by which thresholds and peak amplitudes and latencies are measured, which is time-consuming and prone to user variability. Automated methods have offered promising alternatives for ABR classification, but they have sometimes been limited in accuracy or robustness. Here, we developed and tested a supervised convolutional neural network (CNN) based ABR peak classifier that works across sound levels and sound frequencies that can be run quickly on a personal computer using single or dual-channel ABR inputs. For ABR peaks I, III, IV, and V, the classifier achieved over 95% accuracy. High accuracy was maintained even after noise-exposure causing temporary or permanent threshold shifts, and over 90% of peaks were within 0.041 ms (1 sample) of the manually identified peak. Only a few hundred samples were needed to train the network, making it widely amenable to smaller data studies or where the number of subjects or sessions may be low.

Verbal fluency (VF) tasks are widely used to differentiate patients with cognitive impairment from healthy controls, but total word count produced during these tasks becomes unreliable when patients and controls exhibit comparable proficiency. This study examined, in detail, whether item-level and sequential properties of words produced during a VF task could reliably differentiate high-proficiency patients indistinguishable from controls by word count alone. Seventy-seven native Mandarin Chinese speakers (38 controls and 39 patients with mild cognitive impairment or mild dementia) completed a semantic VF task. Participants were subdivided by proficiency into four groups: high-proficiency controls (HC), low-proficiency controls (LC), high-proficiency patients (HP), and low-proficiency patients (LP). The LC and HP subgroups were matched on semantic fluency scores and thus provided a key focus for the investigation. We examined item-level properties (word frequency, contextual diversity, semantic diversity, surprisal) and sequential properties (positional frequency variation) of the words produced. Significant group differences emerged across item-level psycholinguistic properties, though these were primarily driven by the LP group, with no reliable differentiation between LC and HP. Crucially, positional frequency variation distinguished LC from HP. LC participants began their lists with high-frequency words followed by a systematic decline, whereas HP patients produced words within a consistently narrow frequency band throughout. These findings indicate that item-level psycholinguistic properties alone are insufficient to differentiate HP from LC, whereas sequential word frequency variation provides a potential index of cognitive impairment, reflecting underlying differences in semantic retrieval and memory organisation. Future work with larger samples is needed to validate generalisability.