Psychophysiologic Clinic
Perinatal Psychiatry
Infant & Child Psychiatry
Hope Clinic
   Human Brain Indices Reference Database for EEG spectra, ERPs and ERP components:

Human Brain Indices Reference Database

The HBI Database is a revolutionary tool enabling the professional
  • to assess electrical activity of brain systems
  • to construct protocols of individual treatment
  • to monitor intervention effects  
The HBI database was developed to help researchers perform both conventional and quantitative EEG and quantitative ERPs studies. The HBI database is built in WinEEG software but is an add on.  
The HBI database includes the results of processing more than 3000 EEG recordings collected from more than 1000 health subjects with age from 7 to 89 years old. EEGs were recorded at 7 different conditions: Eye Opened, Eye Closed tasks and during performing 5 different tasks using for ERP analysis such as Visual Continuous Performance Task (VCPT), Auditory Task, Reading Task, Mathematical Task and Mismatch Negativity Task.      
The HBI database includes the average spectra, average coherence, average event related potentials (ERPs) and their variance computed for three different referents: linked ears referent, average referent and weighted average referent.
A 19-channel EEG is recorded in two resting conditions with eyes open (minimum 3 minutes), eyes closed (minimum 3 minutes), and five different task conditions, including two stimulus GO/NOGO tasks, arithmetic and reading tasks, auditory recognition and auditory oddball tasks. The characteristics of QEEG are normalized. The mean values and standard deviations for separate age groups are obtained. Deviations from "normality” are assessed by computing Z-scores, standardized measures of deviation of individual EEG parameters from the normative data.

The reference data includes 19-channel EEG recordings in the following groups:

1) children and adolescents of age 7-17 years old – 300 healthy subjects;

2) adults of age 18-60 years old – 500 healthy subjects;

3) elderly people (after age of 61) – 200 subjects.

It also includes recordings of 300 ADHD children and adolescents, as well as numerous recordings in other kind of patients (patients with epilepsy, OCD, addiction, depression, whiplash, etc.)

Inclusion/Exclusion criteria presume an uneventful perinatal period, no head injury with cerebral symptoms, no history of neurological or psychiatric diseases, no convulsions, normal mental and physical development, average or better grades in school.

19-channel EEG was recorded in two resting conditions with eyes open (more than 3 minutes), eyes closed (more than 3 minutes), and four different task conditions, including two stimulus GO/NOGO task (20 minutes), arithmetic and reading tasks (both 20 Minutes) and auditory task.

To reduce amount of time for pre-processing the data some procedures (such as artifact correction and spike detection) are automated. To assess both global and local features of EEG two montages (global average and local average according to Lemos) are used. Absolute amplitude and power spectra, averaged and two-channel coherences, wavelet-transformations and ERPs are computed off-line and mapped into 2D representations or into 3D images using LORETA technology. Dipole approximation methods are provided in addition.

The computed characteristics are normalized (using log transformations). Mean values and standard deviations for separate age groups are obtained. Comparison with the data base consists of computing z-scores – standardized measures of deviation of individual EEG parameters from the reference data.

In addition, ERPs are subjected to independent component analysis. Using this methodology, separate components associated with distinctive psychological operations are extracted. Each component is characterized by time dynamics and topography. Spatial filters are built up on the basis of these topographies and enable the users to extract the amplitude of each component from the individual ERPs. Comparing these amplitudes with the reference data gives the insights concerning different stages of information processing in the individual under assessment.




Event-Related Potentials


Brain responses (i.e. evoked potentials) to psychological tasks are decomposed into independent components. The components are associated with distinctive psychological operations. Comparing the amplitude and latency of the components with the normative data gives new insights into the different stages of information processing in patients.

In clinical settings the HBI Database is a valuable resource for individualized treatment planning. An example of such an application is presented in the figure below. Simply viewing the raw EEG of an ADHD patient (top, left) does not reveal any abnormality; however, compression of the data into spectra and comparing the spectra with the normative data shows statistically significant (p<0.01) deviation from normality in the theta frequency range (top, right spectra) which is reflected in the central areas (see map at bottom). The electromagnetic tomography of the theta activity is presented at the bottom of the Figure.



A Unique Database for ICA Components   

An example of comparing the components of evoked potentials is presented below. The Fig. depicts time dynamics of four different components for an ADHD patient (thin line) in comparison to norms (thick line). The components are associated with comparison operation, attentional shift, engagement, and monitoring operations. The maps of the components for these patients and norms are presented in the middle. Electromagnetic tomographies of the components are presented at the right. Only one component is selectively reduced in this patient, as shown by the red fill-in. Our studies show that the entire ADHD population can be separated into distinct categories, each characterized by selective suppression of a distinct component, and each responding to a specific medication.