Analyze neural, physiological, and behavioral time-series data.
- European Data File Format: Read, write, and explore EDF/EDF+ file format data
- Deep Learning: Apply LSTM networks, 1-D convolutional layers, and CNNs with time-frequency analysis for signal classification and prediction tasks
- Time-Frequency Analysis: Compute short-time Fourier or continuous wavelet transforms and time-varying coherence between signals
- Multiresolution Analysis: Separate signal components with wavelet-based or data-adaptive multiresolution analysis techniques
- Feature Extraction: Automatically extract deep features from time-series data using a wavelet scattering framework
- Signal Labeling: Label signals automatically and interactively, and visualize labeled signals, with the Signal Labeler app
Analyze images, volumes, and videos at the neuron, brain, and subject scales.
- Blockwise Big Image Representation: Process and display N-D images that are too large to fit in memory, including labeled and multiple resolution images, by representation as smaller blocks
- Neuroimaging and Microscopy Data: Access image slices and volumes from 3-D NIfTI and TIFF files
- Volumetric Data: View labeled volumetric data interactively with the Volume Viewer app, extract image slices, including at oblique angles, and apply over 70 process 3-D image processing functions
- Deep Learning: Apply 2D and 3D CNN models for object detection and semantic segmentation, GAN models for image style transfer, and LSTM models for video classification
- Image Labeling: Apply and view labels to image data via ROI objects and the interactive Image Labeler and Video Labeler apps
Create, train, and run predictive models for neuroscience data.
- Deep Learning Customization: Build custom training loops and custom layers more easily with automatic differentiation and nested layers
- Framework Interoperability: Import models and layers from the Tensorflow and Tensorflow-Keras frameworks; and import/export from and to other frameworks via the ONNX model format
- Deep Learning Experiments: Compare networks trained under various conditions with the Experiment Manager app
- Machine Learning: Discover clusters and noise in data with the DBSCAN algorithm and the Cluster Data live task
Create, share, and scale data analyses.
- Graphics: Flexible data distribution plots, including swarm and box charts; and export graphics for use in scientific publications
- Live Editor and Tasks: Create rich documents combining code, text, figures, interactive controls, and animations; and complete interactive tasks for common data analysis operations
- Unlimited Scaling: MATLAB Parallel Server supports unlimited parallel computing for every user on campus
- Cloud Computing: Apply your campus license to run and scale MATLAB in the cloud, connect to remote data sources such as Amazon S3, and run MATLAB in standard or customized containers
Process live signals for brain recordings, behavioral control systems, and BCIs.
- Stateflow: Graphically design state machine logic for behavioral control systems, runnable in MATLAB or Simulink
- MATLAB Coder: Translate over 3500 MATLAB and toolbox functions to ANSI C or C++ code for faster performance and real-time applications
- HDL Coder: Target FPGA hardware for video processing and closed-loop experiments using high-level MATLAB or Simulink programming
- Multithreading: Call MATLAB asynchronously from user-created threads using the C++ engine API
- Performance: Run existing MATLAB code over two times faster