Strategy classification per epoch
Each trajectory segment labeled with strategy type, duration, and spatial features used for classification.
CSVJSON
← Capabilities
Locomotor Pattern Classification
Classify how animals move, not just where
Automated detection of thigmotaxis, center avoidance, wall-hugging, scanning, and directed movement — spatial strategies that carry more information than distance alone.
8
Locomotor strategies classified
33ms
Frame-level strategy assignment
3
Validated paradigms (OF, WM, BM)
0
Manual expert classification needed
The problem
Movement strategy is more informative than total distance but harder to measure
Two animals with identical total distance can show very different spatial strategies — one exploring the center, the other hugging the walls. Strategy classification traditionally requires an expert to manually categorize trajectory segments, limiting throughput.
- Thigmotaxis ratio requires defining wall zones and computing proportional time — a custom analysis per arena
- Strategy transitions within a session (e.g., thigmotaxis → exploration) require temporal segmentation of the trajectory
- Manual strategy classification by expert raters limits throughput to 5-10 sessions per hour
The solution
Trajectory feature extraction with automatic strategy labeling
ConductVision segments trajectories into behavioral epochs and classifies each segment using spatial features: wall proximity, angular velocity, path tortuosity, and directionality. Eight named strategies are assigned automatically.
- Eight strategies: thigmotaxis, wall-hugging, center exploration, scanning, directed movement, circling, immobility, erratic
- Temporal dynamics: see when and how often the animal switches strategies within a session
- Arena-agnostic: works in open field, water maze, Barnes maze, and custom arenas
Endpoints
Locomotor strategy outputs
Strategy summary statistics
Proportion of session time in each strategy, number of strategy transitions, and dominant strategy by time bin.
CSV
Strategy transition matrix
Markov transition probabilities between strategies — which strategy follows which, and how this changes across sessions.
CSVPNG
Applications
Paradigms using locomotor pattern analysis
Open field thigmotaxis quantification
Thigmotaxis ratio is the most widely used open field anxiety measure. Automated strategy classification provides it without custom zone definitions.
Measures
- Thigmotaxis ratio
- Center exploration time
- Strategy transition frequency
Water maze strategy evolution
Track the transition from random search to scanning to directed swimming across training days — a more sensitive learning measure than latency.
Measures
- Strategy type per trial
- Day of strategy transition
- Scanning-to-directed shift
Circling and stereotypy detection
Unilateral lesion models produce characteristic circling patterns. Automated classification detects and quantifies circling without manual observation.
Measures
- Circling frequency
- Circling direction bias
- Circling bout duration
Age-related strategy simplification
Aged animals use fewer strategy types and show reduced strategy switching. Pattern classification reveals cognitive rigidity.
Measures
- Strategy diversity index
- Transition rate by age
- Dominant strategy shift
Compared to typical systems
How ConductVision differs
| Feature | ConductVision | Typical systems |
|---|---|---|
| Strategy classification | Automated, 8 types | Manual expert, 2-3 types |
| Temporal dynamics | Within-session transitions tracked | Single strategy per trial |
| Arena compatibility | Any arena shape | Circular mazes only (water maze) |
| Transition analysis | Markov transition matrix | Not available |
| Throughput | Real-time, unlimited | 5-10 sessions per hour (manual) |
Related capabilities
Zone and Spatial Analysis
Configurable spatial zones with automated dwell time, entry count, and heatmap analysis for any arena.
Heatmap Visualization
Real-time spatial occupancy heatmaps with color-coded density overlays on the arena view.
Precision Tracking
High-resolution 30 fps tracking that captures sub-second behavioral events conventional systems miss.
See movement strategy, not just movement distance
Upload open field or maze recordings for automatic locomotor strategy classification.
