Biology subjects

Cheng, K. Y.

Publications and source records attributed to Cheng, K. Y..

6 recordsLinked to original sources

bioRxiv2026

Deep Representation Learning on Whole-Brain Population Dynamics Uncovers Geometrically Separable Neural Codes

Learning interpretable low-dimensional representations of whole-brain neuronal dynamics remains a major computational challenge in systems neuroscience. We present a wiring-agnostic deep-learning framework that couples a convolutional encoder with a temporal transformer to learn compact representations directly from volumetric calcium imaging of the entire Drosophila melanogaster brain. Trained to classify 16 experimental conditions that factorially combine metabolic state (fed, starved), sensory modality (olfaction, gustation, or combined), and stimulus valence (appetitive, aversive, or conflicting), the model organizes pan-neuronal whole-brain population activity into geometrically distinct, condition-specific clusters. Analysis of the models latent space reveals that state, modality, and valence are encoded along three near-orthogonal axes: a separable structure that emerges from the classification objective without explicit disentanglement constraints. Spatial attribution and regional importance analyses link modality decoding to distinct anatomical circuits, whereas metabolic state and valence related information show weaker regional specificity and broader distribution across the brain. Our approach does not require anatomical annotation, neuronal identification, or connectivity information, and thus provides a scalable foundation for comparative whole-brain imaging and representation learning of brain wide dynamics.

neuroscience↗

bioRxiv2024

Myoelectric Prosthesis Control using Recurrent Convolutional Neural Network Regression Mitigates the Limb Position Effect

Although state-of-the-art myoelectric prostheses offer persons with upper limb amputation extensive movement capabilities, users have not been afforded a reliable means to control common movements required in daily living. Many proposed prosthesis controllers use pattern recognition, a method that learns and recognizes patterns of electromyographic (EMG) signals produced by the users residual limb muscles to predict and execute device movements. Such control becomes unreliable in high limb positions--a problem known as the limb position effect. Pattern recognition often uses a classification algorithm; simple to implement, but limits user-initiated control to only one device movement at a time, at a single speed. To combat position-related control deficiencies and classification controller constraints, we developed and tested two recurrent convolutional neural network (RCNN) pattern recognition-based solutions: (1) an RCNN classification controller that uses EMG plus positional inertial measurement unit (IMU) signals to offer one-speed, sequential movement control; and (2) an RCNN regression controller that uses the same data capture technique to offer simultaneous control of multiple movements and device movement velocity. We assessed both RCNN controllers by comparing them to a commonly used linear discriminant analysis classification controller (LDA-Baseline). Participants without upper limb impairment were recruited to perform multipositional tasks while wearing a simulated prosthesis. Both RCNN classification and regression controllers showed improved functional task performance over LDA-Baseline, in 11 and 38 out of 115 metrics, respectively. This work contributes an RCNN regression-based controller that provides accurate, simultaneous, and proportional movements to EMG-based technologies including prostheses, exoskeletons, and even muscle-activated video games.

bioengineering↗

bioRxiv2023

A multifaceted suite of metrics for comparative myoelectric prosthesis controller research

Upper limb robotic (myoelectric) prostheses are technologically advanced, but challenging to use. In response, substantial research is being done to develop user-specific prosthesis controllers that can predict a persons intended movements. Most studies that test and compare new controllers rely on simple assessment measures such as task scores (e.g., number of objects moved across a barrier) or duration-based measures (e.g., overall task completion time). These assessment measures, however, fail to capture valuable details about: the quality of device arm movements; whether these movements match users intentions; the timing of specific wrist and hand control functions; and users opinions regarding overall device reliability and controller training requirements. In this work, we present a comprehensive and novel suite of myoelectric prosthesis control evaluation metrics that better facilitates analysis of device movement details--spanning measures of task performance, control characteristics, and user experience. As a case example of their use and research viability, we applied these metrics in real-time control experimentation. Here, eight participants without upper limb impairment compared device control offered by a deep learning-based controller (recurrent convolutional neural network-based classification with transfer learning, or RCNN-TL) to that of a commonly used controller (linear discriminant analysis, or LDA). The participants wore a simulated prosthesis and performed complex functional tasks across multiple limb positions. Analysis resulting from our suite of metrics identified 16 instances of a user-facing problem known as the "limb position effect". We determined that RCNN-TL performed the same as or significantly better than LDA in four such problem instances. We also confirmed that transfer learning can minimize user training burden. Overall, this study contributes a multifaceted new suite of control evaluation metrics, along with a guide to their application, for use in research and testing of myoelectric controllers today, and potentially for use in broader rehabilitation technologies of the future.

neuroscience↗

bioRxiv2022

Walking elicits global brain activity in Drosophila

Movement-correlated brain activity has been found across species and brain regions. Here, we used fast whole-brain lightfield imaging in adult Drosophila to investigate the relationship between walk and neuronal activity. We observed brainwide activity that tightly correlated with spontaneous bouts of walk. While imaging specific sets of excitatory, inhibitory, and neuromodulatory neurons highlighted their joint contribution, spatial heterogeneity in forward walk- and turning-induced activity allowed parsing unique responses from subregions and sometimes individual neurons. For example, previously uncharacterized serotonergic neurons were inhibited during walk. While activity onset in some areas preceded walk onset exclusively in spontaneously walking animals, spontaneous and forced walk elicited highly similar activity in most brain regions. These data suggest a major contribution of walk and walk-related sensory or proprioceptive information to brain state. We conclude that walk-related signals induce global but heterogeneous patterns of activity, allowing for local and brain-wide integration of behavioral state into other processes in the brain.

neuroscience↗

bioRxiv2020

Less physical activity and more varied and disrupted sleep is associated with a less favorable metabolic profile in adolescents

BackgroundSleep and physical activity are modifiable behaviors that play an important role in preventing overweight, obesity, and metabolic health problems. Studies of the association between concurrent objective measures of sleep, physical activity, and metabolic risk factors among adolescents are limited. ObjectiveThe aim of the study was to examine the association between metabolic risk factors and objectively measured school day physical activity and sleep duration, quality, onset, and variability in adolescents. Materials and MethodsWe measured one school week of free-living sleep and physical activity with wrist actigraphy in 252 adolescents (146 girls), aged 15.8{+/-}0.3 years. Metabolic risk factors included body mass index, waist circumference, total body and trunk fat percentage, resting blood pressure, and fasting glucose and insulin levels. Multiple linear regression adjusted for sex, parental education, and day length was used to assess associations between metabolic risk factors and sleep and activity parameters. ResultsOn average, participants went to bed at 00:22{+/-}0.88 hours and slept 6.2{+/-}0.7 hours/night, with 0.83{+/-}0.36 hours of awakenings/night. However, night-to-night variability in sleep duration (0.87{+/-}0.57 hours) and bedtime (0.79{+/-}0.58 hours) was considerable. Neither average sleep duration nor mean bedtime was associated with any metabolic risk factors. However, greater night-to-night variability in sleep duration was associated with higher total body ({beta}=1.9{+/-}0.9 %/h, p=0.03) and trunk fat percentage ({beta}=1.6{+/-}0.7 %/h, p=0.02), poorer sleep quality (more hours of awakening) was associated with higher systolic blood pressure ({beta}=4.9{+/-}2.2 mmHg/h, p=0.03), and less physical activity was associated with higher trunk fat percentage (p=0.04) and insulin levels (p=0.01). ConclusionGreater nightly variation in sleep, lower sleep quality, and less physical activity was associated with a less favorable metabolic profile in adolescents. These findings support the idea that, along with an adequate amount of sleep and physical activity, a regular sleep schedule is important to the metabolic health of adolescents.

pathology↗

bioRxiv2018

Olfactory and neuromodulatory signals reverse visual object avoidance to approach in Drosophila

Innate behavioral reactions to sensory stimuli may be subject to modulation by contextual conditions including signals from other modalities. Whereas sensory processing by individual modalities has been well-studied, the cell circuit mechanisms by which signals from different sensory systems are integrated to control behavior remains poorly understood. Here, we provide a new behavioral model to study the mechanisms of multisensory integration. This behavior, which we termed odor-induced visual valence reversal, occurs when the innate avoidance response to a small moving object by flying Drosophila melanogaster is reversed by the presence of an appetitive odor. Instead of steering away from the small object representing an approaching threat, flies begin to steer towards the object in the presence of odor. Odor-induced visual valence reversal occurs rapidly without associative learning and occurs for attractive odors including apple cider vinegar and ethanol, but not for innately aversive benzaldehyde. Optogenetic activation of octopaminergic neurons robustly induces visual valence reversal in the absence of odor, as does optogenetic activation of directional columnar motion detecting neurons that express octopamine receptors. Optogenetic activation of octopamine neurons drives calcium responses in the motion detectors. Taken together, our results implicate a multisensory processing cascade in which appetitive odor activates octopaminergic neuromodulation of visual pathways, which leads to increased visual saliency and the switch from avoidance to approach toward a small visual object.

neuroscience↗