Epilepsy Diagnosis Dataset

information

The Bonn Dataset is one of the most famous and widely used datasets in the field of brain signal processing (EEG), which is specifically used for the diagnosis and prediction of epilepsy.

This dataset was published by the University of Bonn, Germany, under the supervision of Dr. Ralph Andrzejak in 2001 and is known as a standard benchmark for testing machine learning and deep learning algorithms.

Below, we will examine the full details of the structure and features of this dataset:

Dataset structure (5 subsets)

This dataset consists of 5 different subsets (Sets) known as A, B, C, D, E (or sometimes Z, O, N, F, S). Each set contains 100 signal segments.

Set A (or Z):

Subject: Healthy Volunteers.

Condition: Awake, with eyes open.

Recording type: Surface EEG (from the scalp).

Set B (or O):

Subject: Healthy Volunteers.

Condition: Awake, with eyes closed.

Recording type: Surface EEG.

Set C (or N):

Subject: Epileptic patients (in between seizures or Interictal).

Recording location: Intracranial electrode, recorded from the hippocampal formation of the hemisphere opposite the epileptic area.

Condition: Normal brain activity of the patient (without seizure).

Set D (or F):

Subject: Epileptic patients (interictal).

Recording site: Intracranial electrode, recorded exactly from the epileptogenic zone.

Condition: Patient's brain activity (no seizure).

Set E (or S):

Subject: Epileptic patients (ictal).

Recording site: Intracranial electrode.

Condition: Rhythmic seizure activity.