Getting Started:

Installation

To install aopy, run the following command in your terminal (don’t install aopy by mistake, that’s a different package):

> pip install aolab-aopy

Or to update an existing installation:

> pip install aolab-aopy --upgrade

Overview of functions

Module	contents	Examples
data	Directly loading and saving data from bmi3d, peslab, and results	load_ecube_data(), save_hdf()
preproc	Reorganize data into a standard format, largely automated for bmi3d	get_trial_segments(), proc_exp()
precondition	Clean and prepare neural data for users to interact with	downsample(), get_psd_multitaper(), detect_spikes()
postproc	Separating neural features such as LFP bands or spikes binning. And (currently) loading preprocessed data	extract_mtm_features(), get_kinematic_segments()
analysis	Compute firing rates, success rates, direction tuning, etc.	calc_success_rate(), calc_rms()
visualization	Neural data plotting	plot_spatial_map(), plot_raster(), plot_tfr()
utils	Helper functions, math, other things that don’t really pertain to neural data analysis	generate_test_signal(), detect_edges(), derivative()

Supported systems

Currently aopy supports data from aolab BMI3D and pesaran lab wireless data.

Data from experiments comes from several sources (e.g. experiment hdf files, optitrack csv, binary neural data). To manage all these different sources of data, aopy has parsing functions that standardize the format and contents of these files.

BMI3D

A simple example:

/data/raw/
├── hdf/
|   ├── test20210310_08_te1039.hdf
│   └── ...
├── ecube/
|   ├── 2021-03-10_BMI3D_te1039/
│   |   ├── AnalogPanel_32_Channels_int16_2021-03-10_10-03-58
|   |   └── DigitalPanel_64_Channels_bool_masked_uint64_2021-03-10_10-03-58
│   └── ...

import aopy
data_dir = '/data/raw'
result_dir = '/data/preprocessed/beignet'
block = 1039
files = aopy.data.get_filenames_in_dir(data_dir, block)
result_filename = aopy.data.get_exp_filename(block)
aopy.preproc.proc_exp(data_dir, files, result_dir, result_filename)

Once preprocessed, you can inspect the hdf file using aopy.data.get_hdf_dictionary():

preprocessed_te1039.hdf
├── exp_data
│   ├── task
│   ├── state
│   ├── clock
│   ├── events
│   ├── trials
│   └── <raw bmi3d data>
└── exp_metadata
    ├── source_dir
    ├── source_files
    ├── n_cycles
    ├── n_trials
    ├── bmi3d_start_time
    └── <raw bmi3d metadata>

See Preproc: for more details on the data format. To add mocap and spiking data you would call:

aopy.preproc.proc_mocap(data_dir, files, result_dir, result_filename)
aopy.preproc.proc_spikes(data_dir, files, result_dir, result_filename)

The hdf file would now contain:

preprocessed_te1039.hdf
├── exp_data
│   └── ...
├── exp_metadata
│   └── ...
├── mocap_data
│   └── data
├── mocap_metadata
|   ├── samplerate
│   ├── source_dir
|   ├── source_files
|   └── <raw mocap metadata>
├── spikes_data
│   └── ...
└── spikes_metadata
    └── ...

(proc_spikes doesn’t actually exist as of this writing)

To load a single variable from the preprocessed file, use:

trials = aopy.data.load_hdf_data(result_dir, result_filename, 'trials', 'exp_data')

Or to load an entire group:

exp_metadata = aopy.data.load_hdf_group(result_dir, result_filename, 'exp_metadata')

For a more comprehensive example, see the Examples section of this documentation.

Peslab

Documentation in progress.

aopy.data.peslab

Memory Limits

When working on shared resources, e.g. lab servers, please make use of the aopy.utils.memory module to manage hardware memory limits.

See this example for details.