KiMA

KiMA is a free, browser-based gait analysis suite built for researchers studying locomotor recovery after spinal cord injury. It takes pose-estimation data from DeepLabCut and turns it into publication-ready kinematic plots, gait parameters, and statistical summaries — all without installing any software.

Developed by Manojkumar at the Venkatesh Lab, KiMA supports five experiment paradigms: ladder rung walking, foot-fall pattern detection, CatWalk spatial analysis, treadmill locomotion, and multi-animal batch comparison with PCA.

Key Features

No Installation Required

Runs entirely in your browser. Your data stays on your machine — nothing is uploaded to any server.

DeepLabCut Integration

Directly reads CSV files exported from DeepLabCut. Just make sure your body-part labels match the expected naming convention.

Interactive Visualizations

Batch Processing & PCA

Process entire cohorts at once, assign treatment groups, and run principal component analysis to compare across animals.

Publication-Ready Exports

Export plots as SVG/PNG, download gait metrics as CSV, and generate animated gait sequence videos.

Privacy First

All analysis runs client-side in JavaScript. No data leaves your browser — suitable for pre-publication and sensitive datasets.

Analysis Modules

Ladder & Rung

Single-trial hindlimb kinematics from ladder/rung walking. Generates stick figures, joint angles, gait cycle parameters, and animated sequence playback.

Labels: toe, mtp, ankle, knee, hip, iliac crest Open →

Foot Fall

Detect ground contacts for all four paws, visualize contact-peak signals, generate paw-specific 2D trajectory plots and per-paw videos.

Labels: RF, LF, RH, LH (x, y, likelihood) Open →

Cat Walk

Spatial stepping analysis along a walkway. Calculates projected paw-contact maps, per-paw stride length, and fore/hind stride width.

Labels: RF, LF, RH, LH with paw coordinates Open →

Treadmill

Continuous locomotion analysis with stick-figure player, cycle-by-cycle trends, joint-angle distributions, and early/mid/late block comparisons.

Labels: toe, mtp, ankle, knee, hip, iliac crest Open →

Batch & PCA

Process multiple animals in one session, assign treatment groups (e.g., SCI vs. sham), export combined metrics, and run PCA for group-level comparison.

Upload multiple CSVs from any single-trial module Open →

Quick Start

1 Pick the module Choose the tool that matches your experiment type (ladder, foot fall, catwalk, or treadmill).

2 Check your labels Ensure your DeepLabCut CSV has the required body-part names. See the Manual for the full list.

3 Load data or try examples Upload your CSV or click "Load Example Data" inside any module to test the workflow with sample files.

4 Analyze and export Review interactive plots, tune settings if needed, then export figures (SVG/PNG), CSV metrics, or gait videos.

Supported Data Format

KiMA reads CSV files exported from DeepLabCut. The first three rows contain metadata (scorer name, body-part labels, and coordinate axes). Data rows follow with frame-by-frame x/y coordinates and likelihood scores for each tracked body part.

Each module expects specific body-part label names. If your DLC project uses different naming, rename the columns in the CSV header before uploading. Every module includes built-in example data so you can try the full workflow before using your own files.

Before You Start

Correct DLC labels matter. KiMA reads body-part names directly from your CSV. If names don't match expected labels, parsing will fail.

Choose the matching module. Don't use Foot Fall data in Ladder analysis or CatWalk data in Treadmill analysis. Each module expects a different body-part layout.

Each tool has built-in example data so you can test the workflow before loading your own files.

Required Labels

Module	Required Labels
Ladder / Treadmill	`toe`, `mtp`, `ankle`, `knee`, `hip`, `iliac crest`
Foot Fall / Cat Walk	`RF`, `LF`, `RH`, `LH` with `x`, `y`, `likelihood`
PCA	Aggregated metrics CSV from KiMA batch analysis

Ladder & Rung

Required labels: toe, mtp, ankle, knee, hip, iliac crest
What it shows: Interactive stick plot, gait parameters, angle distributions, per-cycle gait metrics, and sequence playback.
Best use: Checking one animal or one trial in detail before going to batch analysis.
How to use it: Upload one CSV, inspect the plot first, then open Gait Parameters and Statistical Analysis, and finally export images or CSV outputs.

Foot Fall

Required labels: RF, LF, RH, LH with x, y, likelihood
What it shows: Foot-fall counts, paw-only 2D plots, contact-peak signals, separate paw videos, and contact tables.
Best use: Simple visual interpretation of paw contact pattern and stance timing.
How to use it: Upload the paw CSV, tune likelihood and movement thresholds if needed, then review counts, contact peaks, and paw videos before exporting results.

Cat Walk

Required labels: RF, LF, RH, LH with paw coordinates
What it calculates: Projected paw-contact map, stride length per paw, and fore/hind stride width.
Best use: Spatial stepping analysis along a walkway or CatWalk-like setup.
How to use it: Upload the CatWalk-style CSV, verify the projected map looks correct, then review the stride length and width plots and export the metrics table.

Treadmill

Required labels: toe, mtp, ankle, knee, hip, iliac crest
What it shows: Treadmill player, cycle summaries, joint-angle distributions, per-cycle trends, and early/mid/late comparisons.
Best use: Continuous locomotor monitoring rather than single-step inspection.
How to use it: Upload the treadmill CSV, review the stick player first, then inspect cycle trends and block comparisons before exporting the cycle table or images.

Single Run

Ladder & Rung

Use this for one hindlimb run from a ladder or rung experiment. This is the main single-sample kinematic workflow.

Required labels: `toe`, `mtp`, `ankle`, `knee`, `hip`, `iliac crest`.
What it shows: interactive stick plot, gait parameters, angle distributions, per-cycle gait metrics, and sequence playback.
Best use: checking one animal or one trial in detail before going to batch analysis.
How to use it: upload one CSV, inspect the plot first, then open Gait Parameters and Statistical Analysis, and finally export images or CSV outputs.

Paw Contact

Foot Fall

Use this for paw-contact analysis when the key question is where and when each paw touches during locomotion.

Required labels: `RF`, `LF`, `RH`, `LH` with `x`, `y`, `likelihood`.
What it shows: foot-fall counts, paw-only 2D plots, contact-peak signals, separate paw videos, and contact tables.
Best use: simple visual interpretation of paw contact pattern and stance timing.
How to use it: upload the paw CSV, tune likelihood and movement thresholds if needed, then review counts, contact peaks, and paw videos before exporting results.

Runway Geometry

Cat Walk

Use this for runway paw-placement analysis when you want stride length and stride width rather than full hindlimb joint kinematics.

Required labels: `RF`, `LF`, `RH`, `LH` with paw coordinates.
What it calculates: projected paw-contact map, stride length per paw, and fore/hind stride width.
Best use: spatial stepping analysis along a walkway or CatWalk-like setup.
How to use it: upload the CatWalk-style CSV, verify the projected map looks correct, then review the stride length and width plots and export the metrics table.

Continuous Run

Treadmill

Use this for continuous treadmill gait where the interest is cycle behavior across time, stability, and trend changes during the run.

Required labels: `toe`, `mtp`, `ankle`, `knee`, `hip`, `iliac crest`.
What it shows: treadmill player, cycle summaries, joint-angle distributions, per-cycle trends, and early/mid/late comparisons.
Best use: continuous locomotor monitoring rather than single-step inspection.
How to use it: upload the treadmill CSV, review the stick player first, then inspect cycle trends and block comparisons before exporting the cycle table or images.

File Upload

Click to upload or drag & drop

CSV files only

Download Example 1 Download Example 2

Plot Styling

Plot Title

X-Axis Label

Y-Axis Label

Plot Background Color

Background Color:

Highlighted Frame Style

Point Size

Line Width

Text Size

Body Part Colors

Toe:

MTP:

Ankle:

Knee:

Hip:

Iliac:

Background / Unselected Frames

Color

Frame Color:

Opacity

10%

Frame Step

Draw one background frame every N steps to reduce clutter.

General Settings

Labels & Title Color

Text Color:

Likelihood Threshold

Download Options (Interactive Plot)

DPI Quality

100600

300 DPI

Background

Image size: ~2400 x 2100 pixels

Interactive Plot

Hover to see frame number. Click to select or deselect.

Upload a CSV file to begin

Gait Parameters

Calculated gait kinematics and joint angles (46 parameters)

Parameter	Value	Description
Upload a CSV file to calculate parameters

Hip Rise Analysis

Quantify vertical hip elevation, peak-to-trough rise amplitude, and frame-by-frame motion in centimeters.

Calibration & Reference Settings

Pixel-to-CM Conversion Ratio (px/cm)

px = 1 cm

Set pixels per 1 cm (5.5 px/cm = ~6 cm rise scale).

Baseline Reference

Defines zero reference line for vertical height.

Signal Filter / Smoothing

Reduces tracking jitter in vertical height.

Hip Rise Cutoff Threshold (cm)

Threshold line & highlight cutoff on plots.

Quick Scale Presets:

Max Elevation

-- cm

Peak physical height

Avg Step Rise

-- cm

Mean peak-to-trough rise

Rise Threshold

1.00 cm

Cutoff threshold line

Above Threshold

-- / --

Cycles meeting threshold

Total Excursion

-- cm

Vertical excursion (Max-Min)

Hip Elevation Trajectory (Line Plot)

Vertical hip position (cm) across frames. Red dashed line indicates your cutoff threshold. High points are highlighted.

Hip Rise Amplitude per Gait Cycle (Bar Plot)

Quantified vertical rise (cm) achieved during each step cycle. Green bars reach or exceed the threshold line.

Step Cycle Breakdown

Cycle #	Start Frame	End Frame	Trough Height (cm)	Peak Height (cm)	Hip Rise (cm)	Threshold Status
Upload CSV to view cycle breakdown

Angle Distribution Analysis

Distribution of joint angles across the **entire dataset**.

Chart Options

Chart Type

Box Plot Violin Plot Violin + Box

Show Data Points

Plot Color

Plot Color:

Text Size

Download Dimensions

Width (in)

Height (in)

DPI

Gait Metrics

Cadence, stride, and swing/stance from foot speed (assumes 30 fps, foot=toe).

#	Start frame	End frame	Step time (s)	Cadence (spm)	Stance %	Swing %	Stride (px)

Sample Stick Viewer

Queue multiple DeepLabCut CSV files, then load each into the interactive stick plot without re-uploading.

Upload CSV files (multiple allowed)

Each file should follow the DeepLabCut CSV format (frame rows after two header lines).

Queued Samples

Use “Load & View” to open a sample inside the Interactive Plot tab.

Need different settings per sample? load in seconds.

No queued samples yet.

Analysis Model Type

Upload CSV files (multiple allowed)

Each file should follow the DeepLabCut CSV format (frame rows after two header lines).

Download Example 1 Download Example 2

Workflow

Choose the **Analysis Model Type** matching your files.
Upload your CSV tracking data files for multiple animals/runs.
Review the calculated metrics in the batch summary table below.
Assign Groups (e.g. Injury, Control) to perform PCA and compare populations.

Cohort Statistical Report

Automatically computes differences between your defined groups (minimum 2 samples per group required).

Gait Metric	Comparison Groups	Statistical Test	Test Statistic	P-Value	Significance

Upload aggregated metrics CSV

Recommended flow: run Multiple Sample Analysis first and continue to PCA from there. Direct upload is for manual use.

Tips

Label samples as Injured, Treated, Control, etc.
PCA is centered and scaled automatically.
Use color separation to confirm treatment recovery.

Sample	Group
Upload aggregated metrics to edit group labels

File Upload

Click to upload Foot Fall CSV

DLC CSV with RF, LF, RH, LH body parts

Download Example CSV

Detection Settings

Likelihood Threshold

Movement Threshold (px/frame)

Minimum Contact Frames

Frames Per Second

Paws To Plot

RF LF RH LH

Colors

Background

Image Export

DPI

Export

Export Results

Exports detected stance/contact windows for all paws

Summary

Simple counts and averages so users can quickly understand the result.

Total Foot Fall Count

One bar per paw showing how many foot falls were detected.

2D Paw-Only Plot

Only the paw traces and detected foot falls are shown, using the full run length as the axis range.

Paw Line Plot With Contact Peaks

Each line shows paw position across frames, with contact markers placed where foot fall is detected.

Separate Paw Videos

Each paw is shown alone on the same axis range to make the foot fall pattern easier to read.

Frame Frame 0

RF Paw

LF Paw

RH Paw

LH Paw

Detected Contacts

Start/end frame, duration, and contact center for every detected stance window.

Paw	Contact #	Start	End	Frames	Duration (s)	Center X	Center Y
Upload a Foot Fall CSV to begin

Limb Symmetry Analysis

Symmetry Index (SI) comparing left vs. right limbs. Perfect symmetry is 0%; positive/negative values indicate side bias.

Forelimbs (LF vs. RF)

Hindlimbs (LH vs. RH)

Multiple Sample Foot Fall Analysis

Batch process multiple animals, assign treatment/group labels, and export combined foot-fall summaries.

Upload Foot Fall CSV files (multiple allowed)

Sample	Group	RF	LF	RH	LH	Total
Upload multiple Foot Fall files to begin

File Upload

Click to upload Treadmill CSV

DLC CSV with toe, mtp, ankle, knee, hip, iliac crest

Download Example CSV

Settings

Likelihood Threshold

Frames Per Second

Cycle Block Count

Colors

Toe

Ankle

Knee

Hip

Background

Image Export

DPI

Export

Export Results

Exports per-cycle treadmill metrics and joint summaries

Summary

Continuous treadmill summary with cycle count, cadence, stride proxy, and toe clearance.

Treadmill Player

Play the tracked hindlimb as a normalized stick figure through the treadmill run.

Frame Frame 0

Joint Angles Over Time

Toe, ankle, knee, and hip behavior through the full treadmill session.

Cycle Trends

Cycle-by-cycle cadence, stride proxy, and toe clearance to show stability or fatigue over time.

Early / Mid / Late Comparison

Compare treadmill performance blocks across the run.

Per-Cycle Metrics

Each row summarizes one detected treadmill cycle.

Cycle	Start	End	Duration (s)	Cadence	Stride	Toe Clearance
Upload a treadmill CSV to begin

File Upload

Click to upload Cat Walk CSV

DLC CSV with rf, lf, rh, lh paw tracks

Download Example CSV

Settings

Likelihood Threshold

Movement Threshold (px/frame)

Minimum Contact Frames

Colors

Export Results

Exports projected contacts, stride lengths, and stride widths

Summary

Runway-projected paw metrics computed from detected contact points.

Projected Paw Contact Map

Contact points are projected into runway length and lateral width coordinates.

Stride Length

Stride length is computed between successive contacts of the same paw along the runway axis.

Stride Width

Stride width is computed as the lateral separation between left and right fore or hind paw contacts.

Stride Metrics

Projected stride length values and paired width values used for the summary plots.

Metric	Group	Index	Value
Upload a Cat Walk CSV to begin

Key Features

No Installation Required

DeepLabCut Integration

Interactive Visualizations

Batch Processing & PCA

Publication-Ready Exports

Privacy First

Analysis Modules

Ladder & Rung

Foot Fall

Cat Walk

Treadmill

Batch & PCA

Quick Start

Supported Data Format

Tools

Ladder & Rung

Foot Fall

Cat Walk

Treadmill

Batch & PCA

Manual

Before You Start

Required Labels

Ladder & Rung

Foot Fall

Cat Walk

Treadmill

Ladder & Rung

Foot Fall

Cat Walk

Treadmill

Ladder & Rung Analysis

File Upload

Plot Styling

Interactive Plot

Stick Sequence Player

Linear Stick Strip

Gait Parameters

Hip Rise Analysis

Calibration & Reference Settings

Hip Elevation Trajectory (Line Plot)

Hip Rise Amplitude per Gait Cycle (Bar Plot)

Step Cycle Breakdown

Angle Distribution Analysis

Chart Options

Gait Metrics

Sample Stick Viewer

Queued Samples

Batch & Multi-Sample Analysis

Batch Comparison Plots

Cohort Statistical Report

PCA Analysis

Foot Fall Analysis

File Upload

Detection Settings

Export Results

Summary

Total Foot Fall Count

2D Paw-Only Plot

Paw Line Plot With Contact Peaks

Separate Paw Videos

RF Paw

LF Paw

RH Paw

LH Paw

Detected Contacts

Limb Symmetry Analysis

Forelimbs (LF vs. RF)

Hindlimbs (LH vs. RH)

Multiple Sample Foot Fall Analysis

Treadmill Analysis

File Upload

Settings

Export Results

Summary

Treadmill Player

Joint Angles Over Time

Cycle Trends