NWB LFP Loader + Behavior/MNI Alignment

MATLAB script documentation

Usage
README.md • Pipeline Scaffold

Load NWB LFP, map macro-wire channels, align behavior, and attach MNI labels.

This script demonstrates a clean starting point for working with NWB-based intracranial LFP recordings. It supports two acquisition modes (raw vs filt), extracts macrowire channels, aligns LFP time to behavioral trial events, and loads a channel table containing MNI/Freesurfer labels.

Data .nwb + .mat
Sampling 4000 Hz raw / 500 Hz filt
Interfaces acquisition / processing:ephys

! Version Lock

Use MATLAB 2024a with MatNWB 2.5.0 for compatibility. If your NWB schema differs, adjust object paths.

Contents

What this script does

Overview

NWB load (raw vs filtered)

Finds *.nwb files and selects either *raw* or *filter*. Loads timestamps + data from the appropriate NWB path.

Timebase handling

For filtered data, timestamps are downsampled by 8 to match 500 Hz. Raw data keeps the native 4 kHz timestamps.

Macro-wire channel selection

Pulls electrode labels and filters to MA_ channels. Also loads location, hemisphere, and a short brain-area name per channel.

Behavior alignment + trial masks

Loads a behavior MAT file with eventTABLE/behavInfo, then finds the LFP sample index for trial start.

Inputs / Expected folder layout

Setup

The script expects two working directories: one containing NWB files, and one containing behavior and MNI MAT files.

project/
├─ nwb/                      # your NWB files
│  ├─ subject01_filter.nwb
│  ├─ subject01_raw.nwb
│  └─ ...
└─ behavior_mni/             # your MAT files
   ├─ *Behavior*.mat         # contains eventTABLE, behavInfo
   ├─ *MNI*.mat              # contains tmpCSV (channel table)
   └─ ...
  • 1

    Point nwbLOC to the directory containing *.nwb, then cd into it.

  • 2

    Set toLOAD to 'filt' (500 Hz) or 'raw' (4000 Hz).

  • 3

    Set the behavior/MNI directory (the second cd(...)) to your MAT location.

Usage

How-To

Minimal run sequence:

%% Requirements
% MATLAB 2024a + MatNWB 2.5.0

%% 1) NWB directory
nwbLOC = 'DIRECTORY FOR NWB FILES';
cd(nwbLOC)

%% 2) Choose what to load
toLOAD = 'filt'; % 'raw'

%% 3) Run the script sections:
% - NWB load
% - channel selection (MA_)
% - behavior load + alignment
% - MNI labels load

When to use toLOAD = 'raw'

You need high temporal resolution (4 kHz), want to run your own filtering, or need to match raw acquisition timing.

When to use toLOAD = 'filt'

You want a lighter dataset (500 Hz) already high-pass filtered (0.5 Hz) and notch-filtered.

Key variables created

Outputs

LFP time + data

LFP_time (timestamps) and LFP_dataD (double voltage data).

Macro-wire channel metadata

MAchan, chanID2, chanHemi, chanSname.

Behavior and alignment

eventTABLE, behavInfo, sampleINDEX (start of Trial 1).

Trial masks

LA_trials, RA_trials, gamble_trials, plus outcome masks (loss/neutral/gain).

Notes & Common Pitfalls

Important
  • !

    Filtered timestamps: The script assumes the NWB timestamps were stored at 4 kHz and must be downsampled by 8 to match 500 Hz filtered data.

  • !

    NWB paths differ: Filtered LFP is pulled from processing/ecephys/LFP, while raw is pulled from acquisition/MacroWireSeries.

  • !

    Naming consistency: Your snippet uses checkIndex and eventTab. Ensure these match your MAT file variables.