Describe the new feature or enhancement
After briefly discussing this with @drammock, opening this to evaluate if there is interest in MNE-community to add support for reading the data from the Neuralynx acquisition systems (https://neuralynx.com/).
These electrophysiology time series are typically recorded from depth electrodes at high sampling rates (32KHz) to detect single-neuron spiking activity in humans. To my familiarity, a researcher typically will not interact with the raw dataset (due to large file sizes), but with a downsampled version of the continuous recording (.ncs) usually in combination with event files (.nev).
Information on the type of data files in a typical Neuralynx dataset:
Example dataset/research paper:
Describe your proposed implementation
In brief, Neuralynx datasets are stored as a single .ncs (continuously sampled) files for each channel. A dataset directory containing only .ncs channel files could like something like this:
>>> ls ./neuralynx_dataset
LACC1.ncs LAM5.ncs LPCC1.ncs LPHG5.ncs
LACC2.ncs LAM6.ncs LPCC2.ncs LPHG6.ncs
LACC3.ncs LAM7.ncs LPCC3.ncs LPHG7.ncs
LACC4.ncs LAM8.ncs LPCC4.ncs LPHG8.ncs
It seems to me that on the user end this could look something like:
raw = mne.io.read_raw_neuralynx("./neuralynx_dataset") # read all channel files and header informationor possibly for individual channel files:
raw_ch1 = mne.io.read_raw_neuralynx_ncs("./neuralynx_dataset/LACC1.ncs") # read single channel file and header informationDescribe possible alternatives
The Neo python package has a NeuralynxIO class intended to be used as interface to read Neuralynx files in python, but offers no functionality to analyze and visualize the data.
Additional context
I have not analyzed such data before, so I don't have a sample dataset at hand nor can give examples a typical workflow (but might in the coming months). I've only ever had to read in the header information of a Neuralynx dataset for purposes of documenting it and noticed this was not possible with MNE-Python.
Describe the new feature or enhancement
After briefly discussing this with @drammock, opening this to evaluate if there is interest in MNE-community to add support for reading the data from the Neuralynx acquisition systems (https://neuralynx.com/).
These electrophysiology time series are typically recorded from depth electrodes at high sampling rates (32KHz) to detect single-neuron spiking activity in humans. To my familiarity, a researcher typically will not interact with the raw dataset (due to large file sizes), but with a downsampled version of the continuous recording (
.ncs) usually in combination with event files (.nev).Information on the type of data files in a typical Neuralynx dataset:
Example dataset/research paper:
Describe your proposed implementation
In brief, Neuralynx datasets are stored as a single
.ncs(continuously sampled) files for each channel. A dataset directory containing only.ncschannel files could like something like this:It seems to me that on the user end this could look something like:
or possibly for individual channel files:
Describe possible alternatives
The Neo python package has a NeuralynxIO class intended to be used as interface to read Neuralynx files in python, but offers no functionality to analyze and visualize the data.
Additional context
I have not analyzed such data before, so I don't have a sample dataset at hand nor can give examples a typical workflow (but might in the coming months). I've only ever had to read in the header information of a Neuralynx dataset for purposes of documenting it and noticed this was not possible with MNE-Python.