← Capabilities

Motion Energy Quantification

Continuous activity measurement below tracking resolution

Frame-differencing motion energy captures grooming, tremor, twitching, and other in-place behaviors that centroid-based tracking misses entirely.

Motion Energy Quantification
30fps
Temporal resolution
1px
Spatial sensitivity
0
Tracking model required
Any
ROI or whole-frame analysis
The problem

Centroid tracking misses in-place behaviors

Coordinate-based tracking measures locomotion — distance traveled, velocity, zone entries. But many important behaviors happen in place: grooming, tremor, head bobbing, convulsions. The centroid barely moves during these events, so tracking systems report the animal as "stationary."

  • Grooming produces vigorous limb and head movement with near-zero centroid displacement
  • Tremor and seizure-related movements are invisible to position tracking
  • Freezing (true immobility) and in-place grooming look identical in coordinate data
The solution

Pixel-change motion energy complements coordinate tracking

ConductVision computes frame-to-frame pixel intensity differences across the entire frame or within defined ROIs. The resulting continuous motion energy signal captures all movement regardless of whether the animal changes position.

  • Continuous activity signal at frame rate — every movement registered, no centroid dependency
  • ROI-specific motion energy: isolate head motion, body motion, or limb motion independently
  • Distinguishes true freezing (zero motion energy) from stationary grooming (high motion energy, zero locomotion)
Endpoints

Motion energy outputs

Frame-level motion energy trace

Frame-level motion energy trace

Continuous motion energy value per frame — total pixel change normalized by animal area. Plotted as a time series for visual inspection.

CSVJSON
ROI-specific motion energy

ROI-specific motion energy

Separate motion energy traces for user-defined ROIs — e.g., head region, body region, tail region — enabling body-part-specific activity quantification.

CSV
Activity state classification

Activity state classification

Thresholded motion energy assigns frames to activity states: immobile, in-place active, or locomoting — resolving the freezing/grooming ambiguity.

CSV
Applications

Applications for motion energy analysis

Seizure detection

Convulsion quantification

Seizure events produce characteristic high-amplitude motion energy spikes. Automated detection replaces manual Racine scale scoring.

Measures
  • Convulsion energy magnitude
  • Seizure bout duration
  • Seizure frequency
Grooming

Grooming bout detection

High motion energy with low centroid displacement identifies grooming bouts. Duration and frequency are stress-sensitive measures.

Measures
  • Grooming bout count
  • Total grooming time
  • Grooming intensity
Sleep/wake

Sleep-wake state estimation

Sustained low motion energy indicates sleep. Brief motion energy bursts during low-activity periods indicate micro-arousals.

Measures
  • Sleep bout duration
  • Wake bout duration
  • Micro-arousal frequency
Tremor

Parkinsonian tremor detection

Resting tremor produces low-amplitude, rhythmic motion energy oscillations detectable in frequency domain analysis.

Measures
  • Tremor frequency
  • Tremor amplitude
  • Tremor bout duration
Compared to typical systems

How ConductVision differs

FeatureConductVisionTypical systems
In-place movement detectionYes — pixel-level sensitivityNot detected (centroid-only)
Tracking model requiredNo — model-free analysisRequires trained detection model
ROI-specific analysisConfigurable body regionsWhole-body only
Freezing vs. groomingDistinguished by motion energyBoth classified as stationary
Temporal resolutionPer-frame (33 ms)Per-second or per-bin

Related capabilities

Behavior Classification

Behavior Classification

ML-powered frame-level behavioral classification with temporal smoothing and custom classifier training.

Precision Tracking

Precision Tracking

High-resolution 30 fps tracking that captures sub-second behavioral events conventional systems miss.

Unsupervised Discovery

Unsupervised Discovery

Unsupervised behavioral clustering reveals structure without predefined ethograms — UMAP visualization and novel behavior flagging.

Detect the behaviors your tracker cannot see

Upload a recording and compare motion energy to coordinate-based activity — see what you have been missing.

Request a demoSchedule a plan