Skip to main content
Capability

Analyze & Visualize Outcomes

Cortex transforms raw sensor data into meaningful daily behavioral features through an open-source pipeline. The outputs have been validated across dozens of peer-reviewed studies.

The Cortex Pipeline

Cortex is mindLAMP's open-source Python library for turning raw sensor and activity data into research-ready behavioral features. It processes data through three stages, each adding interpretive value. Cortex documentation →

Raw Features
32 Modules
  • GPS coordinates
  • Accelerometer (x, y, z)
  • Screen state & battery
  • Step counts
  • Call & text metadata
  • Survey responses
  • Cognitive game results
→
Primary Features
6 Modules
  • Significant locations (GPS clustering)
  • Trips (origin, destination, duration)
  • Screen-on bouts
  • Acceleration jerk magnitude
  • Game scores & response times
  • Survey question-level scoring
→
Secondary Features
23 Modules
  • Home time, entropy, trip distance
  • Screen duration, step count
  • Call & text frequency
  • Sleep duration
  • Data quality score
  • One row per participant per day

From Raw Data to Behavioral Insights

A week of raw sensor data becomes a compact set of behavioral features that reveal patterns no single data stream could show alone.

Thousands of GPS readings become a single daily "home time" value. Accelerometer streams become step counts. Screen events become usage duration. Researchers run Cortex on their data to extract these features across participants and time periods.

~86,400 GPS readings
→
hometime, entropy, trip_distance
~432,000 accelerometer samples
→
jerk_magnitude
144 GPS coverage bins
→
data_quality
~40 screen on/off events
→
screen_duration
Call/SMS metadata events
→
call_frequency, text_frequency

Full feature reference →

In Practice

Recent peer-reviewed research using Cortex analysis tools.

Schizophrenia Clinical Subtypes
Integrated cognitive, functional, and digital phenotyping assessments to identify clinical subtypes in schizophrenia, demonstrating how Cortex features complement traditional measures.
Byun et al., 2025 — Molecular Psychiatry
View project →
Mount Sinai Mood & Anxiety
Used digital measures of physical activity and screen time to predict real-world changes in depression and anxiety symptoms, linking passive sensing to clinical outcomes.
Beltrán et al., 2025 — NPJ Digital Medicine
View project →
SHARP-NIMHANS Cognitive Study
Analyzed longitudinal digital phenotyping correlates of cognitive performance in schizophrenia spectrum disorders across a 12-month study.
Das et al., 2025 — Schizophrenia Research
View project →

Data Access

Access data programmatically in the language you already use. SDK documentation →

Python
pip install lamp-cortex
R (via reticulate)
Use Python SDK from R
JavaScript
npm i lamp-core
REST API
api.lamp.digital/v1
JSONata
Built-in query language for filtering data

Dive Deeper

Explore the technical documentation for Cortex, data access, and analysis tools.

Cortex Getting Started
Install Cortex, authenticate, and run your first feature extraction.
Read guide →
Feature Reference
Complete list of raw, primary, and secondary features with input/output schemas.
View reference →
API & SDKs
Python and JavaScript SDKs, REST API, and data access workflows.
Read guide →
Data Portal
Browser-based data querying and CSV export for researchers who prefer not to use code.
Read guide →