_load_results(files_df; generate_Xs::Bool = true)

Internal function to load Unfold models into memory. Can also be used to load data after file information was loaded lazily (lazy=true) using load_results()

add_event_files!(layoutDF)

Function to find and load all events file-paths into Layout-DataFrame.

bids_coeftable(model_df)

Turns all models found in model_df into tydy DataFrames and aggregates them in a new DataFrame.

bids_effects(model_df::DataFrame, effects_dict::Dict)

Calculate mariginalized effect on all subjects found in the model dataframe using effects_dict.

bids_layout(bidsPath::AbstractString;
    derivatives::Bool=true,
    specific_folder::Union{Nothing,AbstractString}=nothing,
    exclude_folder::Union{Nothing,AbstractString}=nothing,
    ses::Union{Nothing,AbstractString}=nothing,
    task::Union{Nothing,AbstractString}=nothing,
    run::Union{Nothing,AbstractString}=nothing)

Main function to load paths of all subjects in one bids_root folder. Will return a DataFrame containing all found paths with specific subject information. Used before loading data into memore using load_bids_eeg_data

Keywords

• derivatives::Bool = true
  Look for data in the derivatives folder
• specific_folder::Union{Nothing,AbstractString} = nothing
  Specify a specific folder name in either derivatives or bids_root to look for data.
• exclude_folder::Union{Nothing,AbstractString} = nothing
  Exclude a specific folder from data detection.
• ses:Union{Nothing,AbstractString} = nothing
  Which session to load; loads all if nothing
• task::Union{Nothing,AbstractString} = nothing
  Which task to load; loads all if nothing
• run::Union{Nothing,AbstractString} = nothing
  Which run to load; loads all if nothing

check_df(files_df, ses, task, run)

Internal; Checks if the multiple sessions/task/runs are found if none of these are provided

extract_subject_id!(files_df, file)

Internal function to get subject information from dataframe.

list_all_paths(path)

Internal function to find pathfiles

load_bids_eeg_data(layout_df; verbose::Bool=true, kwargs...)

Load data found with bids_layout into memory.

• loading_function::Union{Nothing,Function} = nothing
  Function to load EEG data; if nothing provided and PyMNE.jl is loaded, it will use that as default. The function must take a single argument: the file path to the EEG data file.
• verbose::Bool = true
  Show ProgressBar
• kwargs...
  kwargs for CSV.read to load events from .tsv file; e.g. to specify delimeter

load_events(layoutDF::DataFrame; kwargs...)

Internal function to load events based on paths in the layout Df

load_results(bids_root::String;
    derivatives_subfolder::String="Unfold",
    lazy::Bool=false,
    generate_Xs::Bool = true,
    ses::Union{Nothing,AbstractString}=nothing,
    task::Union{Nothing,AbstractString}=nothing,
    run::Union{Nothing,AbstractString}=nothing)

Load Unfold models existing in a derivatives_subfolder in your BIDS root folder.

Keywords

• derivatives_subfolder::String = "Unfold"
  Defines in which subfolder of bids_root/derivatives to look for Unfold models.
• lazy::Bool = false
  Do not actually load the dataset into memore if true, only return a dataframe with paths
• generate_Xs::Bool = true
  By default recreate the designmatrix; Can be set to false, to improve loading time.
• ses::Union{Nothing,AbstractString} = nothing
  Which session to load; loads all if nothing
• task::Union{Nothing,AbstractString} = nothing
  Which task to load; loads all if nothing
• run::Union{Nothing,AbstractString} = nothing
  Which run to load; loads all if nothing

rename_to_latency(data_df)

This is a convenience function to add a :latency collumn (needed by Unfold) based on another variable in the events_df (e.g. sample)

run_unfold(data_df, bf_vec; 
	remove_time_expanded_Xs=true, 
	extract_data = raw_to_data, 
	verbose::Bool=true, 
	kwargs...)

Run Unfold analysis on all data in data_df. Needs to have PyMNE.jl loaded!

Arguments

• data_df::DataFrame
  DataFrame containing BIDS data as returned by load_bids_data(). Must contain collumns: :subject, :ses, :task, :run, :raw, :events
• bf_vec
  Basis function vector as expected by Unfold.jl's fit() function.
  

Can be one of: \
- `["eventname" => (formula, basisfunction)]` for overlap corrected models
- `["eventname" => (formula, timewindow)]` for mass-univariate models, where indicates (start, stop) in seconds with `typeof(timewindow) = Tuple{Real, Real}`

Keywords

• remove_time_expanded_Xs::Bool = true
  Removes the timeexpanded designmatrix which significantly reduces the memory-consumption. This Xs is rarely needed, but can be recovered (look into the Unfold.load function)
• extract_data::function = nothing
  Specify the function to extract the data to a data array. Falls back to raw_to_data which uses get_data and allows to pick channels if no function is povided AND PyMNE is loaded - see @Ref(raw_to_data).
• verbose::Bool = true)
  Show ProgressBar or not.
• kwargs...
  Will be passed to both the fit() and extract_data() calls as function inputs.
  For possible kwargs to fit() please have a look at the Unfold.jl API: https://unfoldtoolbox.github.io/UnfoldDocs/Unfold.jl/stable/references/functions/

save_results(results::DataFrame, bids_root::String; 
    derivatives_subfolder::String="Unfold",
    overwrite::Bool=false)

Function to save unfold models in your BIDS root folder. Automatically creates a derivatives_subfolder (default = "Unfold") in the derivatives and subsequentely safes each model in results according to BIDS. Example of path to saved file: bids_root/derivatives/Unfold/sub-XXX/eeg/sub-XXX_ses-XX_task-XXX_run-XX_unfold.jld2

Keywords

• derivatives_subfolder::String = "Unfold"
  Creates the named subfolder and saves Unfold models according to BIDS.
• overwrite::Bool = false
  Does not overwrite existing datasets; can be set to true.

unpack_events(df::DataFrame)

Unpack events into tidy data frame; useful with AlgebraOfGraphics.jl

df is expected to be a UnfoldBIDS DataFrame where events are loaded already.

unpack_results(results_df)

Unpack all results into one tidy dataframe/ coeftable.