Skip to content

LLAMAS Data Structure

jalevitt edited this page Apr 4, 2023 · 3 revisions

THE LLAMAS DATA STRUCTURE


When you save your recording in LLAMAS, it will save a .mat file with the variable 'EEG', which is a struct containing a large number of fields and sub-fields. It is worth discussing what some of these are, as it may otherwise be opaque.

EEG_var

Here are some of the more pertinent fields of this variable:

Recording - This field stores the downsampled EEG data in column vectors. The last column will be timestamps (note that timestamps may be unreliably due to unresolved bugs - in practice you should infer timestamps from your sample rate) and if you are using triggers, the penultimate column will contain your triggers. All other columns will contain EEG data. Units for EEG channels will generally be in uV, although this could depend upon how you are feeding data into LSL.

Recording_orig - This contains zero-padded data at your native (non-downsampled) sample rate. Because this is takes up tons of memory, only the last couple minutes of data are stored here. If you need access to the full recording at the native sample rate, please see the .txt file which is created automatically with your recording.

Kalman_Signal - This contains your Kalman filtered data, assuming you have selected 'Kalman_chunk' from the setting menu.

KalmanTarget - Is a vector of channel indices, which indicate which channels are being Kalman filtered.

KalmanRegressors - Is a vector of which channels are being used as references for the Kalman filter.

ChunkSizes - This is record of how many sample each chunk contained at the downsampled frequency as it was received from LSL.

Latency - This is a vector showing how much time elapsed between the receipt of a chunk and the completion of processing of that chunk. Units in seconds.

KeyPresses - This stores keyboard presses during the recording. This capability has not been thoroughly tested, use at your own risk.

vars - This is a struct that contains a large number of fields that are mostly useful for backend development.

PostProcessing - This cell array contains the names of any postprocessing functions used during this recording.

Clone this wiki locally