DONALD WLODKOWIC LAB

AUTOMATED ANIMAL TRAINING

Investigating proto-cognitive processes such as associative learning, spatial memory, and decision-making requires experimental paradigms in which stimuli are delivered with precise timing, and — crucially — as a direct consequence of specific animal behaviours. These closed-loop conditioning paradigms have been standard in mammalian research for decades, but accessible and affordable systems for small aquatic organisms have been absent, creating a significant methodological gap in cognitive research.

Our lab has addressed this gap through the development of two complementary open-source platforms. The first, NeuroBox, is a low-cost hardware and open-source software interface built on an Arduino microcontroller that enables fully programmable, automated delivery of up to three independent stimuli during behavioural biotests. The system was validated using tap-startle escape responses in larval zebrafish, producing consistent and reproducible fear responses across multiple stimulation cycles.

The second platform, TrackingBot, is a real-time animal tracking application that couples on-the-fly video analysis with closed-loop hardware control. Rather than delivering stimuli on a fixed schedule, TrackingBot detects an animal's position in real time and triggers external devices the moment a defined behavioural condition is met — the fundamental requirement for operant conditioning.

Together, these tools open new experimental avenues for asking whether simple organisms can form associations, remember spatial locations, and adapt their behaviour based on experience — questions that sit at the heart of proto-cognition research.

Bai, Y., Henry, J., Cheng, E., Perry, S., Mawdsley, D., Wong, B.B.M., Kaslin, J., & Wlodkowic, D. (2024). Toward real-time animal tracking with integrated stimulus control for automated conditioning in aquatic eco-neurotoxicology. Environmental Science & Technology. https://doi.org/10.1021/acs.est.3c07013  

Bai, Y., Henry, J., Kreuder, F., Mawdsley, D., Kaslin, J., & Wlodkowic, D. (2024). An open-source programmable interface for sensory–motor biotests with zebrafish (Danio rerio). Zebrafish. https://doi.org/10.1089/zeb.2023.0067 

ELECTROTHERMAL PELTIER PLATFORMS

Analysing thermotactic behaviours presents significant technical challenges: stable thermal gradients must be precisely generated and spatiotemporally controlled, and existing shuttle-box systems designed for this purpose are large, expensive, and poorly suited to small aquatic invertebrates and larval vertebrates.

Our lab has developed two successive generations of miniaturised thermoelectric platforms to address this gap. The first-generation system used solid-state Peltier elements mounted at opposing ends of compact PMMA test chambers to generate self-balancing thermal gradients across six parallel chambers, validated between 20.5°C and 39°C and stable for up to 60 minutes. This proof-of-concept platform demonstrated the feasibility of high-throughput thermal preference biotests across nine aquatic species and was fabricated at a fraction of the cost of conventional shuttle-box systems.

The second-generation platform represents a substantial engineering advance. It integrates programmable 40W Peltier assemblies with closed-loop microcontroller feedback, enabling precise, independently controlled heating and cooling across shallow low-volume PMMA chambers (77×13×6mm). The system supports both binary thermal zones and continuous linear gradients within the same hardware, switchable on demand, with dynamic zone-inversion achievable within approximately three minutes and minimal thermal drift. Gradient stability was validated over 72 hours of continuous operation (±0.3°C), and temperature distributions were confirmed to be unaffected by the presence of freely swimming animals.

Henry, J., Bai, Y., Kreuder, F., Saaristo, M., Kaslin, J., & Wlodkowic, D. (2022). A miniaturized electrothermal array for rapid analysis of temperature preference behaviors in ecology and ecotoxicology. Environmental Pollution 314, 120202. https://doi.org/10.1016/j.envpol.2022.120202 

Han, X., Kumari, S., Do, H., Bevan, C., Wasielewski, O., Hall, M.D., & Wlodkowic, D. (2026). Thermotactic decision-making in aquatic invertebrates: high-resolution behavioral analysis of ecotoxicological effects. Environmental Science & Technology. https://doi.org/10.1021/acs.est.5c16058 

MICROFLUIDICS & CHEMOTAXIS

Studying chemosensory behaviours in small aquatic invertebrates requires precise spatial and temporal control of chemical gradients — a capability that conventional test systems do not readily provide. To address this, we developed a scalable millifluidic perfusion platform built from optically transparent PMMA, in which peristaltic pumps drive laminar flow through microchannel arrays to create stable binary fluid zones within circular test chambers. CFD simulations were used to optimise chamber geometry and validate flow conditions before fabrication, and zone formation was confirmed experimentally within two minutes of actuation. The modular chipboard architecture allows throughput to be increased simply by adding parallel units, with up to 12 chambers imaged simultaneously by a single camera.

As a proof-of-concept, the platform was validated using chemotaxis assays with marine amphipods Allorchestes compressa, demonstrating its capacity to resolve rapid avoidance responses with quantitative precision.

Bai, Y., Henry, J., & Wlodkowic, D. (2020). Chemosensory avoidance behaviors of marine amphipods Allorchestes compressa revealed using a millifluidic perfusion technology. Biomicrofluidics 14, 014110. https://doi.org/10.1063/1.5131187 

ANALYSIS OF PHOTOTACTIC BEHAVIOURS

Quantitative analysis of phototactic behaviours in small aquatic invertebrates requires precise, independently controllable light stimuli and imaging systems capable of resolving individual animals across many chambers simultaneously — a combination not available in commercial behavioural analysis platforms. To address this, we developed a purpose-built, low-cost system integrating a custom 24-well PMMA test plate, a concentric pinhole LED array that confines the photic stimulus to a discrete 2 mm zone at the centre of each chamber, and an orthogonal infrared illumination system that spatially separates the imaging and stimulus light paths. This decoupling — absent in all existing commercial systems — enables the creation of sharply defined, custom-configured light and dark zones without artefacts in the video record. Stimulus delivery is controlled via in-house C# software that allows full programmability of light intensity, duration, and ON/OFF cycling. Up to 48 chambers and 240 animals can be imaged simultaneously by a single IR camera, with automated high-throughput tracking and a dedicated bioinformatic analysis pipeline. Total system cost is under USD 1,000.

The platform's sensitivity was demonstrated in validation experiments with Artemia franciscana nauplii, which revealed not only canonical phototactic responses but also a previously undescribed phenomenon: persistent light searching behaviours (LSBs) following extinguishing of the stimulus — active seeking of the extinguished light zone that the tightly controlled pinhole design made detectable for the first time.

Henry, J., Bai, Y., Williams, D., Logozzo, A., Ford, A., & Wlodkowic, D. (2022). Impact of test chamber design on spontaneous behavioral responses of model crustacean zooplankton Artemia franciscana. Lab Animal https://doi.org/10.1038/s41684-021-00908-7 

Bai, Y., Henry, J., Karpiński, T.M., & Wlodkowic, D. (2022). High-throughput phototactic ecotoxicity biotests with nauplii of Artemia franciscana. Toxics 10, 508. https://doi.org/10.3390/toxics10090508 

HIGH-THROUGHPUT BEHAVIOURAL TESTS

Quantitative chemobehavioural phenotyping of small aquatic organisms requires careful optimisation of the entire analytical pipeline — from video acquisition through to data extraction — to generate reproducible, unbiased endpoints at scale. We have systematically characterised how video file parameters affect tracking accuracy, demonstrating that frame rate transcoding introduces irreparable data loss, while carefully optimised compression and background subtraction techniques can substantially improve both processing speed and detection fidelity without sacrificing quantitative accuracy. Benchmarking of three animal tracking software packages — Ethovision XT, ToxTrac, and LoliTrack — against identical datasets revealed significant inter-algorithm variability, establishing that software selection is a non-trivial source of experimental bias in high-throughput behavioural studies.

These findings inform the design of our custom analytical platforms, which integrate purpose-built PMMA test chambers, wide-field IR imaging, and high-throughput batch processing pipelines to enable simultaneous recording and tracking of large cohorts across multiple chambers without motorised stages or temporal bias. Platform sensitivity was validated across phototaxis and locomotion assays with Artemia franciscana nauplii.

Henry, J., Rodriguez, A., & Wlodkowic, D. (2019). Impact of digital video analytics on accuracy of chemobehavioural phenotyping in aquatic toxicology. PeerJ 7, e7367. https://doi.org/10.7717/peerj.7367 

Henry, J., & Wlodkowic, D. (2020). High-throughput animal tracking in chemobehavioral phenotyping: current limitations and future perspectives. Behavioural Processes 180, 104226. https://doi.org/10.1016/j.beproc.2020.104226 

Bai, Y., Henry, J., Karpiński, T.M., & Wlodkowic, D. (2022). High-throughput phototactic ecotoxicity biotests with nauplii of Artemia franciscana. Toxics 10, 508. https://doi.org/10.3390/toxics10090508 

LIVING EMBRYO ARRAYS

Conventional multiwell plate formats are poorly suited to high-throughput chemobehavioural analysis of zebrafish embryos: they do not support precise in-test positioning, shadow-free illumination, or temporal consistency across large cohorts. To address these limitations, we developed two distinct custom living embryo array platforms.

The first-generation system was a 3D multilayer PMMA Lab-on-a-Chip device featuring 21 miniaturised embryo traps arranged in linear arrays across 12 independent clusters, dimensioned to the footprint of a standard 96-well plate. Embryo trapping exploited combined gravitational sedimentation and low-pressure suction at the trap base, achieving 100% trapping efficiency.

The second platform scaled substantially to a 189-trap static PMMA array (130×60 mm), enabling simultaneous in-test positioning and wide-field imaging of embryo stereotypic behaviours. A custom-built 4K infrared camera system — converted to the 850 nm spectrum and mounted on a vibration-free column — enabled simultaneous video imaging of 189 embryos during photomotor response (PMR) bioassays in zebrafish embryos at 28–30 hpf.

Zhu, F., Wigh, A., Friedrich, T., Devaux, A., Bony, S., Nugegoda, D., Kaslin, J., & Wlodkowic, D. (2015). Automated Lab-on-a-Chip technology for fish embryo toxicity tests performed under continuous microperfusion (μFET). Environmental Science & Technology 49, 14570–14578. https://doi.org/10.1021/acs.est.5b03838 

Henry, J., Bai, Y., Kreuder, F., Mawdsley, D., Kaslin, J., & Wlodkowic, D. (2022). Accelerating chemobehavioral phenotypic screening in neurotoxicology using a living embryo array system. Zebrafish 19, 32–35. https://doi.org/10.1089/zeb.2021.0072. 

EMBRYO PHOTOMOTOR RESPONSE (PMR)

The photomotor response (PMR) bioassay — in which zebrafish embryos at 24–36 hours post-fertilisation are stimulated with a high-intensity light pulse, inducing a transient increase in the frequency of body flexions — offers a rapid, reproducible window into early neuromodulatory function. Despite its advantages, widespread implementation has been constrained by the lack of a straightforward, high-throughput bioinformatic approach for analysing behavioural data at scale.

We developed a dedicated bioinformatic pipeline for rapid PMR analysis of large embryo cohorts on pre-recorded high-definition video files. The workflow integrates digital post-processing of native video files (grayscale conversion, de-noising, and sharpening), ingestion into custom Ethovision XT templates configured for activity-based pixel intensity detection, and automated batch analysis across up to 100 arenas per video file — enabling simultaneous processing of 200 embryos per recording. 

PMR responses are partitioned into pre-excitation, excitation, and refractory phases, with area-under-the-curve analysis and behavioural heatmap barcoding providing both statistical outputs and rapid visual summaries across chemical concentration series. The paper also provides a systematic analysis of pixel density constraints inherent to wide-field imaging at 1080p resolution, and identifies 4K and higher sensor configurations as the practical path toward improved per-embryo tracking fidelity without sacrificing analytical throughput.

Henry, J., Bai, Y., Kreuder, F., Mawdsley, D., Kaslin, J., & Wlodkowic, D. (2024). A bioinformatic protocol for rapid analysis of zebrafish embryo photo-motory responses (PMR) in neurotoxicity testing. Comparative Biochemistry and Physiology, Part C 277, 109833. https://doi.org/10.1016/j.cbpc.2024.109833.