Skip to content

Latest commit

 

History

History
156 lines (131 loc) · 9.02 KB

README.md

File metadata and controls

156 lines (131 loc) · 9.02 KB
Raw


Logo

Mondragon-Gonzalez et al., 2024, Nature Neuroscience

Closedloop TriFilterNet

TriFilterNet utilizes a feature selection and supervised classification approach to process electrophysiological recordings, identifying pre-grooming events within ongoing behavior. The system employs a set of triangular filters to reduce the size of the feature set, yielding a matrix of decorrelated coefficients. Subsequently, a minimal-architecture artificial neural network is used to classify behavioral states and detect pre-grooming instances based on these coefficients.
Link to article »

Table of Contents

1 System Requirements

  • MATLAB version: The code was developped using Matlab R2017b and also tested using Matlab R2019b.
  • Operating System: The code was developped and tested under Windows 7 and 10.
  • MATLAB Toolboxes: Deep Learning Toolbox
  • Third-party MATLAB Packages: wrapped functions of RHD2000 MATLAB RHD file reader ( link to documentation) and Kamil Wojcicki (2024). HTK MFCC MATLAB, Mathworks exchange link , MATLAB Central File Exchange.

2 Installation Guide

  1. Clone the repository to your local machine: 
    git clone https://github.com/LizbethMG/Mondragon_NatNeuro_CL.git
  2. Navigate to the cloned directory: 
    cd Mondragon_NatNeuro_CL
  3. Download the data used for the `dataset1/m1` and add it to your project's folder data/m1.
  4. Add the project folder to your MATLAB path. You can do this in two ways:
    5. Using the MATLAB Command Window
    • Open MATLAB.
    • In the Command Window, type the following command, replacing yourpath with the path to your project: 
      addpath('yourpath');
savepath;
    Using the MATLAB Set Path Dialog
    • Open MATLAB.
    • On the Home tab, in the Environment section, click Set Path.
    • Click Add with Subfolders.
    • Browse to the project folder and click OK.
    • Click Save and then Close to save the changes.
  5. You haven't installed Matlab's Deep Learning Toolbox yet, go to Matlab's main menu >> Adds-On>> Get Adds-on>> Search for "Deep Learning Toolbox" >> Install
Typical install time: 10 min.

3 Description of Code's Functionality


→ Please refer to Figure 4 of the article as it provides a detailed schematic description of the code's workflow. The complete description can be found in the Methods section.
♦ For more detailed information, users can refer to the in-code comments within the script and its functions. The core functionality of the main script:

INPUT

  • m1_config.m configuration file of the experiment
  • data/ m1/ data files

Configuration and initialization

  • continous_classif__settings.mlx settings for data processing

Data processing

  • continuous_classif_fct.m data processing is completely done in this function, including:

Acquisition and segmentation

  • slmg_nextFromRecordedData.m retrieves the next data segment from a continuous ephys recording file
  • slmg_read_Intan_files.m read ephys files in a folder
  • slmg_read_Intan_file.m read ephys from a single file

Feature decomposition

  • slmg_power.m computes the power from a signal sample.
  • slmg_mfcc.m computes cepstral coefficients from a signal.

Supervised learning

  • slmg_NNCompute.m classification using an artificial neural network (single electrode)
  • slmg_ClassNNdecision.m classification prediction according to single electrode classification results

OUTPUT

  • TimestampsPredict pre-grooming classification timestamps

♦ Expected run time for demo: For a 10-minute experiment (sampled at 30 kHz, 14 channels) like the one on the demo (m1), on a general-purpose computer expect the following approximately runtimes:
~ 8 min -- CPU: Intel Core i7-6700hq @ 2.6 GHz, RAM: 16 GB, Operating System: Windows 10 64-bit, Matlab 2019b.

4 Instructions for Use

Working with the demo data

  • Ensure you added the project folder to your MATLAB path and set it as the current folder. In Matlab Command Window type pwd the result should be 'YOUR_PATH\NatureNeuroscience_2024_TriFilterNet'
  • Run the Live Script file continous_classif__settings.mlx. It will run the code for the demo data (Dataset1/m1) for 10 minutes and return the predictions made in that time window.

Working with Dataset 1

Dataset 1 features the electrophysiological recordings and the associated grooming markers used in the paper to asses the closed-loop algorithm. The data is available at the Open Science Framework DOI: 10.17605/OSF.IO/KDMJT

  • Download dataset 1.
  • Add the downloaded folder to your MATLAB path containing the code, inside 4_data Each sub-folder in the data set 1 corresponds to a single experiment. Each experiment has its configuration file and raw data.
  • In the Live Script file continous_classif__settings.mlx simply change the path to the experimental data and the corresponding configuration files.
    Example:
    • continuous_classif_settings.mlx >> subjects_lists = "m1" to analyze data for experiment m1
    • continuous_classif_settings.mlx >> subjects_lists = ["m1", "m2", "m3"] to analyze data for experiment m1 to m3

5 Citation

If you use this code or data we kindly ask you to cite our work.

  • Data:

(APA style) Mondragón-González, S. L. (2024, March 26). 2024_Mondragon-Gonzalez_NatureNeuroscience. https://doi.org/10.17605/OSF.IO/KDMJT

  • Article: Mondragon et al 2024: DOI link

@article{Mondragon2024, title = {Closed-loop recruitment of striatal interneurons prevents compulsive-like grooming behaviours}, author = {Sirenia Lizbeth Mondragón-González and Christiane Schreiweis and Eric Burguière}, journal = {Nature Neuroscience}, year = {2024}, url = { https://www.nature.com/articles/s41593-024-01633-3 }}

6 License

Logo <NatureNeuroscience_2024_TriFilterNet> Copyright (C) <2024> <Sirenia Lizbeth Mondragón-González and Eric Burguière>

This program is free software: you can redistribute it and/or modify
it under the terms of the GNU General Public License as published by
the Free Software Foundation, either version 3 of the License, or
(at your option) any later version.

This program is distributed in the hope that it will be useful,
but WITHOUT ANY WARRANTY; without even the implied warranty of
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
GNU General Public License for more details.

You should have received a copy of the GNU General Public License
along with this program.  If not, see <https://www.gnu.org/licenses/>.