Representing an ongoing transcranial direct current stimulation (tDCS) in a tsv stimulus file

Quick overview of different kinds of tES

Transcranial electrical stimulation (tES) is a non-invasive brain stimulation technique used in cognitive neuroscience to produce transient modifications of brain activity. It usually consists in the seamless application of weak direct electrical currents (typically ranging, from 0.5 up to 3 mA) for a few minutes (~10 minutes, usually up to 30 minutes) via a pair of electrodes positioned on the scalp. tES applications include tDCS (transcranial Direct Current Stimulation), tACS (transcranial Alternating Current Stimulation), and tRNS (transcranial Random Noise Stimulation), among others.

tDCS current shape: Continuous, unidirectional (DC) current. Waveform: Direct current with a constant amplitude Polarity: The current flows consistently from the anode to the cathode, with no oscillation or variation in current intensity over time
tACS current shape: Sinusoidal alternating current. Waveform: Oscillating current with a sine wave shape, where the current alternates between positive and negative phases at a specific frequency. Frequency Range: Typically ranges from 1 Hz to 1000 Hz. Polarity: The current continuously alternates direction, creating a rhythmic pattern.
tRNS current shape: Random, high-frequency alternating current Waveform: Stochastic (random) signal with no fixed frequency, covering a broad range of frequencies (usually between 0.1 Hz and 1000 Hz). Polarity: The current rapidly alternates direction in a random pattern.

tDCS in [1]

tDCS application session here starts and ends with current transient ramps (~10 seconds), but for its largest part (~10 minutes) it carries a steady current flow or at an alternate current for tACS or tRNS. The stimulation constitutes a kind of “constant” substrate of the whole experiment, conceptually similar to spatial (eg. electrode placement) or technical (eg. device specifications) features, rather than a periodically presented variation of the conditions, alternating between active (“on-”, starting from an onset time) and silent (“off-”, after a duration time) periods. This does not hold for every tDCS experiment.

The problem

Brain Imaging Data Structure representation of a task is based on the identification of a sequence of events occurred during the experiment and the systematic enumeration in a tsv file, according for example to their order of presentation to the participants (eg. “behavioral experiments”) or to their onset and duration (eg. “task events”).

Several non-invasive brain stimulation methods follow this kind of trend of alternating “on-” and “off-” periods, starting from specific onset times, during or before / after a task execution. That is why including NIBS information within the task- file makes sense.

For tDCS in [1], seizing onsets and durations of the stimulus and is represented by 1 row in the events table, with 1 onset (usually at time = 0) and a duration which is supposed to last for the whole duration of the task. Task itself, however, may not carry a fixed, well-defined duration, therefore indicating 600 (for 10 minutes) or 900 (for 15 minutes) could be misleading. The reader expects the stimulation to start at a specific time and end after the duration, but the actual event sequence could last more.

Solution 1 - tDCS as part of the task

In this solution, tES would be included in the task sequence. In a situation similar to the one described in [1] - experiment 1 - “before” the task.tsv would be like the one in Table 1 (top). Here the cognitive stimulation (audiobook) and the NIBS start at the same time and last for the same time. In this case it is easy to insert the description of tES because the file itself is concise and there is no actual characterization of the task itself (i.e. specific columns, etc). In Table 1 there is described a situation similar to [1] - experiment 1 - “Experimental task”, where participants were instructed to perform a visual task. The very first row is dedicated to the tES characterization and all the other rows represent the actual task. The columns related to the NIBS are completely different from those related to the visual task.

(1) sub-01/ses-audiobook/sub-01_ses-audiobook_task-audiobook_beh.tsv
onset   duration    trial   typology intensity
0   900 audiobook   N/A N/A
0   900 tDCS Real 3

(2) sub-01/ses-experiment/sub-01_ses-experiment_task-experiment_beh.tsv

Onset Duration trial position novelty old color typology intensity
0 900 N/A N/A N/A N/A N/A Sham 3
1.8173 0.8006 1 right old 6 G N/A N/A
4.3358 0.7506 2 left new N/A Y N/A N/A 
6.7877 0.6505 3 right old 3 B N/A N/A
...

Table 1 NIBS description included within the task representation files.

Solution 1 - Pros

unified approach, similar to the one of basically all the other NIBS NIBS occurrence is explicitly mentioned within the task representation, avoiding potential loss of information and bad characterization of the experimental environment

Solution 1 - Cons

tsv file is not uniform. There is a single row carrying only NIBS information and the remaining part of the file describes only a behavioral experiment. there are two independent groups of columns representing NIBS and behavioral experiment. unrelated columns are populated with redundant “N/A” values which lead to a tsv file hard to browse

Solution 2 - tDCS described as part of the stimulation environment

In this solution, tES would be included in the scans.tsv file summarizing the session. In a situation similar to the one described in [1] - experiment 1 - “before” the task.tsv would be like the one in Solution 1 but without the characterization of the NIBS. The cognitive stimulation (audiobook) is described within the scan file, as shown in Table 2, 1 and 2. In Table 2 3 and 4 there is described a situation similar to [1] - experiment 1 - “Experimental task”, where participants were instructed to perform a visual task. Here the scans file includes the characterization of the NIBS. In the behavioral task file there is described the visual task, with no additional rows and a more consistent choice of the columns.

ses-audiobook - “before”
(1) sub-03/ses-audiobook/sub-03_ses-audiobook_scans.tsv
filename tDCS typology intensity
sub-03/ses-audiobook/beh/sub-03_ses-audiobook_task-audiobook_beh.json Real C1 3

(2) sub-03/ses-audiobook/sub-03_ses-audiobook_task-audiobook_beh.tsv
onset   duration    trial   
0   900 audiobook

ses-experiment
(3) sub-03/ses-experiment/sub-03_ses-experiment_scans.tsv
filename tDCS typology intensity
sub-03/ses-experiment/beh/sub-03_ses-experiment_task-experiment_beh.jsonReal C1 3

(4) sub-03/ses-experiment/sub-03_ses-experiment_task-experiment_beh.tsv
Onset Duration trial position novelty old color
1.8173 0.8006 1 right old 6 G
4.3358 0.7506 2 left new N/A  Y
6.7877 0.6505 3 right old 3 B
...

Table 2 NIBS description included in the scans description files, leaving the task description files to the plain behavioral task.

Solution 2 - Pros

tsv files representing the progress of the tasks look consistent and easy to browse. The N/A are present only in those cells related to specific task situations. NIBS information can be enumerated within the scan sequence file facilitating the identification of those experiment with or without NIBS

Solution 2 - Cons

NIBS information is not immediately available within the task representation file. Browsing only the last level of the directory structure might lead to losses of information this might introduce an experiment - specific solution for NIBS representation which might go against the BIDS 80-20 rule of thumb Discussion For the situation described in [1], which is common in tDCS experiments, Solution 2 seem to be the best one, splitting independent pieces of information in separate, conceptually valid boxes. This would also keep a likely human - read file easy to browse and understand. However, this might introduce a way to arrange information which holds mainly for a single, specific technique. In addition to this, recent developments in the technique itself suggest that more

Bibliography

[1] Grasso, P.A., Tonolli, E. & Miniussi, C. Effects of different transcranial direct current stimulation protocols on visuo-spatial contextual learning formation: evidence of homeostatic regulatory mechanisms. Sci Rep 10, 4622 (2020). https://doi.org/10.1038/s41598-020-61626-7

[2] Haslacher, David et al. Working memory enhancement using real-time phase-tuned transcranial alternating current stimulation Brain Stimulation: Basic, Translational, and Clinical Research in Neuromodulation, Volume 17, Issue 4, 850 - 859

viacovella / nibs-bids-sandbox