A dataset of EEG and EOG from an auditory EOG-based communication system for patients in locked-in state

Abstract

The dataset presented here contains recordings of electroencephalogram (EEG) and electrooculogram (EOG) from four advanced locked-in state (LIS) patients suffering from ALS (amyotrophic lateral sclerosis). These patients could no longer use commercial eye-trackers, but they could still move their eyes and used the remnant oculomotor activity to select letters to form words and sentences using a novel auditory communication system. Data were recorded from four patients during a variable range of visits (from 2 to 10), each visit comprised of 3.22 ± 1.21 days and consisted of 5.57 ± 2.61 sessions recorded per day. The patients performed a succession of different sessions, namely, Training, Feedback, Copy spelling, and Free spelling. The dataset provides an insight into the progression of ALS and presents a valuable opportunity to design and improve assistive and alternative communication technologies and brain-computer interfaces. It might also help redefine the course of progression in ALS, thereby improving clinical judgement and treatment.

Fig. 1

The procedure performed during a single day. The figure depicts the sequence of the types of sessions performed by patients and the criteria to progress from one type of session to the next. The patients first performed the Training sessions during which the patient learned to move his/her eyes to generate the signal to control the auditory communication system. At the end of the Training session, a classification model was built, and when the accuracy of the built model was greater than 75% the patients performed the feedback session. During the feedback sessions the patients were provided the feedback of their response, i.e., whether their answer was classified as “yes” or “no”. When the feedback accuracy exceeded 75% the patients first performed a copy speller session and then a free speller during which they could spell whatever they desired.

Fig. 2

Different types of trials in the study. (a) Paradigm describing the sequence of events and sequence of the triggers’ labels used in a single trial for the Training and Feedback sessions. In these types of sessions, 20 questions with “yes” and “no” answers, known by the patient, are presented in a pseudo-random order. (b) Paradigm describing the sequence of events and sequence of the triggers’ labels used during a single trial for the Copy and Free spelling sessions. In these sessions, instead of questions, the patient is presented with options that allow him/her to navigate through his/her predetermined spelling scheme (e.g., sectors, letters). For both spelling sessions, the limit in the number of trials depends only on the patient’s attempts to spell the given target (i.e., Copy speller sessions) or her/his desired sentence (i.e., Open speller sessions). For any type of session recorded, the recording’s start and end are indicated by an “S 9” and an “S 15” trigger.

Fig. 3

EOG and EEG setup. (a) Montage for the minimum number of EOG channels for each recorded session, using the locations LO1 (left cantus) and LO2 (right cantus) for horizontal eye movement, and SO1 (above superior orbit) and IO1 (below inferior orbit) for vertical eye movement. We used the labels EOGL, EOGR, EOGU, and EOGD, respectively, for the online study. (b) Montage for the minimum number of EEG electrodes for each recorded session, emphasizing the central motor (C4, Cz, C3) and prefrontal areas. In this latter case, the location of used electrodes might vary between F3 and F4, or Af4 and Af3. Nevertheless, the total number of electrodes might vary between days of the visits due to the patient’s wellness conditions. The exact number of electrodes and labels used can be verified in the Online-only Tables S1–S4.

Fig. 4

Raw data folder structure. Structure of nested folders containing the raw recordings of the study. According to the patient identifier, the upper level is the folder, which can be PN₁ = 11,13,15 or 16. In the next level, VN₂ indicates the total number of visits available for that patient, and inside it, DN3 indicates the number of days that the visit lasted. Each day’s folder stores subfolders for the Training, Feedback, and spelling (that stores recordings from both the Copy and Free speller sessions). Each of these folders contains a set of files that are the outcome of a recorded session (detailed in the section Data Records), times the number that particular type of session (i.e., n₁, n₂, and n₃) was respectively performed during the day.

Fig. 5

Matlab data fields structure. Nested structure elements containing the values and features of recordings from the study. According to the patient identifier, the upper level is the main structure, which can be PN₁ = 11,13,15 or 16. In the next level, VN₂ indicates the total number of visits available for that patient, and inside it, DN3 indicates the number of days that the visit lasted. Inside each day, there are structures for the Training and Feedback sessions and the spelling sessions (containing recordings from both the Copy and Free speller sessions). Each of these contains a set of structures that result from exporting the *.vhdr raw files for each recorded session, times the number of that particular session type (i.e., n₁, n₂, and n₃) was performed during the day. Read the Data Records section for details on the data exporting.

Fig. 6

Example of a Matlab structure of the data using P11’s data. The figure illustrates the data structure using P11’s data from visit V06 and day D03. (a) Indicates the selection of patient variables and the data fields corresponding to a particular visit and day and inside it, the type and number of sessions performed on the given day. (b) Depicts the presence of different fields upon selecting a session type, in this case, the number of Feedback sessions performed by P11, upon selection of Field named Feedback, and their different elements, as shown in the figure. Read the Data Records section for the detailed description.