Design and Construction of an EEG Data Acquisition System for Measurement of Auditory Evoked Potentials

JOSE ANTONIO GUTIERREZ GNECCHI, LUIS ROGELIO SORIANO LARA AND JULIO CESAR HERRERA GARCIA

Instituto Tecnológico de Morelia, Departamento de Ingeniería Electrónica. Avenida Tecnológico 1500 Col. Lomas de Santiaguito, Morelia, Michocacan, Mexico.

angugi98@netscape.net

Abstract

In order to improve the chances of recovery of

young patients suffering from hypoacusia, it is necessary to conduct hearing diagnostic tests during the first 3 to 6 moths after birth. However, for instance, in rural areas of Mexico where there is little or null access to specialized equipment, it is common to find that patients are not diagnosed until the patients’ diminished hearing capacity has impaired their speech development as well as their and social and school integration. This paper presents the design and construction of the EEG-ITM03 EEG auditory evoked potential data acquisition system intended for diagnostic of hypoacusia. The EEG data acquisition system was tested on 5 healthy young adults and the results were compared to those obtained using a commercial equipment (CADWELL 7200). The results show that the early brain stem evoked potential latencies, related to the hearing process can be detected, even when the system is operated in non-ideal locations for conducting hearing tests. Thus the results suggest that the equipment can be used in clinics without special facilities (i.e. sound proof rooms) as part of routine diagnostic activities.

1. Introduction

Hypoacusis (or hypocusya), refers to the level of hearing impairment of patients. One of the main factors that influence the recovery of patients suffering from hypoacusis is the early detection of auditory pathologies [1]. In particular for newborn patients, it is very important to obtain a diagnosis during the first three to six months after birth, so as to increase the chances of successful recovery and favor speech development. More than 90% of children suffering from moderate or acute hypoacusis are likely to go

through correct hearing, intellectual and emotional development if they are diagnosed during the first year after birth [2].

In particular in rural areas in Mexico, where there is little or null access to diagnostic equipment, it is common to find patients suffering hypoacusia that are not diagnosed until much later in life precluding their integration to social and school life. In the Michoacán State, (Mexico) The Human Communication Institute (Spanish: Instituto de la Comunicación Humana) has reported that 20.13% of the population between 4 and 80 years old show some level of hypoacusia; 4.71% of the population suffer from moderate to severe hypoacusia [3]. There are a number of methods for diagnosis of hypoacusia; otoacoustic emission (EOAE) and impedance audiometry are amongst the most commonly used methods [4]. Another technique that can be used to assess the hearing ability of patients consists of measuring brain activity due to external acoustic stimuli. Brain activity is measured in a non-invasive manner by placing electrodes on the patient’s scalp [5][6]; the resulting data is known as encephalogram (EEG) [7].

Although a great deal of research effort has been put into developing working Brain Machine interfaces [8], still, development of EEG diagnostic equipment occupies an important place in research and development. Improvements and new devices are continually reported and registered for measuring brain stem evoked potentials[9][10][11][12]. Measurement of Event Evoked Potentials (AEP) due to external stimuli, allows the analysis of brain signal processing activities [13]. Thus, the use of EEG measurement equipment with Evoked Potentials analysis capabilities can be a cost-effective solution for the assessment of brain activity due to external auditory stimuli. In particular for newborn patients who can not provide feedback for diagnosis, the objective and non-invasive

Electronics, Robotics and Automotive Mechanics Conference 2008

978-0-7695-3320-9/08 $25.00 © 2008 IEEE

DOI 10.1109/CERMA.2008.86

547

nature of the technique can provide useful information for early diagnosis of hypoacusia.

This paper presents the design and construction of EEG measurement equipment with auditory evoked potential analysis capabilities on request from the City Health Office (Spanish: Direccion de Salud Municipal) of Morelia, Mexico. The aim is to produce equipment that can be used to asses the hearing capabilities of patients even if the study is carried out under non-controlled conditions (i.e noise proof facilities). Such equipment could then be used in locations where sound proof facilities are not available and a quite room with ambient noise may suffice. The EEG equipment is initially intended for being used with a host PC; the software must be intuitive, provide the analysis functions commonly encountered in commercial equipment and permit registration of patient data. 2. EEG-ITM03 EEG Data Acquisition System

Figure 1 shows the block diagram of the EEG-ITM03 Auditory Evoked Response data acquisition system.

Figure 1. The EEG-ITM03 Auditory evoked response consist of A) a three electrode scheme for measuring EEG signals, B)

isolation power supply and coupling amplifier, C) signal conditioning and digitizing section, D) auditory stimuli generation section and E)

communication with a host PC for programming (JTAG) and data transferring.

The equipment consists of five main sections. The

first part has an analogue instrumentation amplifier, filter and signal conditioning for a implementing a three-electrode measurement strategy similar to other EEG devices [14] (Figure 1A). The resulting signal is transferred to the digitizing part of the circuit by means

of an isolation amplifier (Figure 1B). A further signal conditioning and filter section feeds the analogue to digital converter of the MSP430F149 microcontroller used for generating the auditory stimuli and acquiring the EEG measured signals (Figure 1C). Pulses (clicks) generated with the microcontroller are isolated and amplified to provide the auditory stimuli (Figure 1D) through an earphone.

The EEG-ITM03 includes a JTAG connection for in-system microcontroller programming and an RS232 interface port for transferring the acquired data. The RS232 connection was tested successfully with USB-RS232 interface connections for use with modern computers that do not include an RS232 port. The equipment can be operated by a 9V battery or by connection to the mains. Figure 2 shows the EEG-ITM03 with electrodes attached and the controls in the frontal panel.

Figure 2. EEG-ITM03 EEG data acquisition

system for auditory evoked potential measurements

2.1 Acoustic stimulus

One of the most commonly used methods for

generating the auditory stimuli for auditory evoked potential tests consists of producing a sequence of pulses to drive a set of earphones, and record the resulting brain electrical activity over a period of a few milliseconds. Since the evoked response has a

548

magnitude of just a few microvolts, the process is repeated 1000 to 2000 times and the results are averaged to improve the Signal-To-Noise Ratio. Assuming that the resulting data is a function, only, of the auditory stimuli, the averaged signal represents the hearing process. The data acquisition process has to be synchronized with the application of auditory stimuli, which consists of a 0.2 miliseconds pulse, which in turn drives the earphones. The pulse is generated 6.66 times per second. The output signal was calibrated using a TES1350 decibel meter, and a graduated scale is provided behind the amplitude control potentiometer in the frontal panel of the equipment.

2.2 Measurement of auditory evoked potentials

AEPs are characterized by three parameters:

polarity, latency (i. e. the moment of peak occurrence after stimulus presentation) and scalp distribution. Figure 3 shows a typical reference wave pattern for diagnosis of AEP.

Figure 3. A) Typical AEP showing the three components and B) diagnostic reference values extracted from a test report sheet (Courtesy of Clinica de Especialidades de

Morelia). There are three types of components: early latencies

or components (up to 10 miliseconds after the stimulus has been applied), middle components (from 10 miliseconds to 50 miliseconds) and late components (after 50 miliseconds).

Although measurement of auditory evoked potentials does not constitute a hearing test per se, it can provide information about the hearing process. In addition, evoked potentials are not influenced by the state of consciousness (i. e. patient can be asleep) and result in objective data that do not require feedback from the patient. The occurrence of waves I - VII

(Figure 3B) can be indicative of hearing, whereas absence of some (or all) of them may indicate a hearing impairment. To produce the AEP graph it is necessary to apply the stimulus for 1000 to 2000 epochs (i.e. application of the stimulus and subsequently measure the resulting EEG signals) (1):

k kS nμ= + (1) Where Sk represents the kth epoch, μ is the

deterministic evoked response and nk represents zero-mean white noise uncorrelated to μ [15]. Averaging over N data sets, YN, can then be written as (2):

1 1

1 1N N

N k kk k

Y S nN N

μ− −

= = +∑ ∑ (2)

The EEG-ITM03 uses a three-electrode scheme for

measuring EEG activity. The positive electrode is connected ipsilateral to the acoustic stimulation. The negative electrode is connected contralateral to the acoustic stimulation. The reference electrode is placed in the forehead. An instrumentation amplifier with CMMR (Common Mode Rejection Ratio) better than 110 dB is used for measuring the EEG signals. Measuring the EEG activity is synchronized with the stimulation process. The Code Composer software was used for the writing the microcontroller program. 2 miliseconds worth of data are acquired prior to producing the auditory stimulus; 15 miliseconds worth of data (after the stimulus is applied) are acquired by the microcontroller at a rate in excess of 47 ksps (kilo samples per second) with 12 bit resolution. The data acquisition process results in a set of 800 12-bit values which are averaged with the next data set, to produce the EEG auditory evoked potential measurement. The number of times the process is repeated is controlled by the C++ program residing in the host computer.

2.3 Electrical Safety considerations

Although the circuitry uses an isolated power

supply and isolation amplifiers, electrical safety is of great concern since the main purpose is directed towards evaluating hearing of newborns. Before the equipment is tested on patients, measurements were taken under different single-fault and normal operating conditions. The equipment was considered safe if, at least, minimal NFPA 99 leakage current specifications are met: A.- Patient to Ground (isolated): ≤10μA (GND intact) B.- Patient to Ground (isolated): ≤50μA (GND open) C.- Between Leads (isolated): ≤10μA (GND intact) D.- Between Leads (isolated): ≤50μA (GND open) E.- Between Leads (non-isolated): ≤50μA (GND intact or open).

549

The electrical leakage currents were measured as described elsewhere for a prototype EEG [14].

For ground-intact tests, the largest leakage current measured was 7.82 μA, between the reference electrode terminal and the mains ground. For ground-open tests the largest leakage current registered was 14.45 μA between the reference electrode terminal and the mains ground. Both measurements are within the safety specification values (10μA and 50μA respectively) and thus, pending corroboration from a certified laboratory, the equipment was considered safe.

2.4 Software

A program, written in C++, was developed for

controlling the operation of the EEG-ITM03. The program permits to register patient and test information (patient data, type of test, etc.). The program also accepts the operating parameters (i. e. number of times the stimuli will be applied, start and stop the test) and transfers them to the microcontroller prior to each test.

Figure 4. Screen for registering patient and

test information. Once the test has been completed, the program

receives the data from the microcontroller. The EEG data can then be stored, plotted and subjected to further filtering and analysis. The program consists of three main windows. The first window registers the test and patient data (Figure 4). The second window (Figure 5) can be accessed by selecting the “CALIBRATE” button. Selecting the “GO ON-LINE” button displays on-line EEG data for the purpose of offset and gain adjustment of the measured signal.

Figure 5. Calibration window

Activating the “GO TO TEST” initiates the test.

The third window shows the test results (Figure 6).

Figure 6. Test results window. Results are

shown for a healthy 34-years old male patient. The latencies are detected in agreement with

expected values for healthy patients. The test results window can display the results for

both ears. In addition, to help the identification of the latencies (i.e. when the evoked potentials occur), there are buttons for calculating the local maximum values for both signals. A third unprocessed signal can be displayed for comparison. Additional to the averaging function, a low-pass filter can be applied to the resulting evoked potential results. Horizontal displacement bars allow the exact identification of the time when the latencies occur.

550

3.0 Validation of the EEG-ITM03 equipment

In order to validate the performance the EEG-

ITM03, a series of experiments were conducted at the Clinica De Especialidades de Morelia. Commercial equipment (CADWELL 7200) was used for obtaining reference information. The tests were conducted on 5 unmedicated young adults (average age of 25 years old) using a 95 dB acoustic stimulus, at a rate of 6 pulses per second for 2000 epochs. The test procedure was conducted in a sound-proof room, with the patients lying down, eyes closed, to avoid registering data from other types of sensory stimulation. Considering that the proper conditions for conducting the auditory evoked potential tests are not available in the locations where the equipment is intended to be used, the tests where repeated on the same 5 patients on a different location.

Figure 7. Results for patient 1, using the CADWELL 7200 equipment.

In the case of the EEG-ITM03, the tests were

conducted in the Electronics Engineering Laboratory at Morelia Institute of Technology. A quiet location was chosen similar to that than can be obtained in medical facilities dependant of the City Health Office, and rural areas. Figure 7 shows the results for patient 1 obtained under near-ideal conditions with the CADWELL equipment at the Clinica de Especialidades de Morelia.

Figure 8 shows the results obtained for patient 1 obtained with the EEG-ITM03 in the laboratory. The equipment was adjusted to produce 95 dB auditory

stimuli at a rate of 6.66 repetitions per second during 2000 epochs.

Figure 8. Test results using the EEG-ITM03 on

patient 1.

Similar tests were conducted on the other 4 patients under the same non-ideal conditions.

4.0 Results and Discussion

Table 1 shows a summary of the results for the 5

patients tested.

Table 1. Comparison of test results using the CADWELL 7200 equipment and the EEG-ITM03

system for auditory evoked potential measurements. The table shows the values of selected latencies for evaluating the hearing

capacity of patients

The results show agreement between the

CADWELL 7200 and EEG-ITM03 equipments, thus suggesting that the EEG-ITM03 can be used for diagnostic of some forms of hypoacusia, in non-ideal environments. Undoubtedly, there will be differences in the results between the CADWELL and EEG-ITM03 equipments. Different specifications for both devices and testing under different conditions result in the differences shown. However, the results indicate that it is possible to detect the auditory evoked potentials even if the test is conducted on a non-ideal environment. The largest error obtained was for latency V, for patient 5 (7.8%). Still, the result can be used to verify the hearing capacity of the patient. The equipment described in this work is not intended to replace the diagnostic ability of the specialist, but to

551

facilitate the analysis of test results. The software interface was intended to be intuitive and easy to use. The software was developed in C++ for this application; although the use of powerful commercial signal processing software is commonly reported (i. e. MATLAB), this work intends to produce a stand-alone royalty-free device that can easily be upgraded, for including further signal processing algorithms and for distribution amongst the regional medical community. The equipment permits to register the test results and thus record statistical data related to the population hearing conditions that has not been done to date.

5.0 Conclusions and Future Work

Throughout this work the design and construction

of an EEG data acquisition system for detection of auditory evoked potentials was described. The equipment was tested on 5 young adult volunteers to verify its operation. Experiments were conducted in a non-ideal environment, and the results agree with those obtained with commercial equipment in a sound proof room, close to ideal conditions. The availability of the EEG-ITM03 can therefore contribute to early diagnosis of some forms of hypoacusia to improve the chances of recovery of young patients. Current work involves obtaining the medical equipment safety certificate to initiate the experimental work on newborns and contribute to facilitate their language and speech development.

6.0 Acknowledgements

The authors are grateful to Dr. Octavio Ibarra Bravo

from Clinica de Especialidades de Morelia, for his useful suggestions, expertise, and allowing access to the CADWELL equipment.

7.0 References

[1] National Institute of Health: Early identification of Hearing impairment in infants and young children. NIH Consensus Statement, No. 11, 1993, pp. 1-24. [2] Bielecki I., Świetliński J., Zygan L and Horbulewicz A. “Hearing assessment in infants from the hypoacusia risk group”. Med. Sci. Monit., Vol.10 (Suppl 2), 2004, pp. 115-117. [3] Rodríguez Díaz J. A., Chavira Contreras C. L, Montes de Oca Fernández E. “Frecuencia de defectos auditivos en 16 estados de México”, An. Otorrinolaringol. Mex. Vol. 46 No. 3, Junio.-Agosto. 2001, pp. 115-117.

[4] White K. R., Vohr B. R., Behrens T. R. “Universal newborn hearing screening using transient evoked otoacustic emission: Results from the Rhode Island hearing assessment project”, Sem Hear, Vol. 14, 1993, pp. 18-29. [5] S. J. Luck, “An Introduction to the event-related potential technique”, The MIT Press, 2005. [6] T. C. Handy, “Event-Related Potentials: A Methods Handbook", The MIT Press, 2004. [7] N. Schaul. "The fundamental neural mechanisms of electroencephalography", Electroencephalography and clinical neurophysiology Vol. 106, 1998, pp. 101-107. [8] Sajda P., Müller K. R., and Shenoy K. V. “From the Guest Editors”. IEEE Signal Processing Magazine - Special Section - Brain Computer Interfaces. Vol 25, No. 1, Jan 2008, pp. 16-18. [9] Fadem K. C. “Evoked response testing system for neurological disorders”. US Patent Application. US 11/570630. 2005. [10] Köpke W. “Device for Determining Acoustically Evoked Brainstem Potentials”. US Patent. US 7197350 B2. 2007. [11] Givens G., Balch D. C., Murphy T., Blanarovich A., Keller P. Systems, “Methods and products For diagnostic Hearing Assesments Distributed Via the use of a Computer Network”. US 6916291 B2. 2005. [12] DeCharms R. C. Methods for Measurement and Analysis of Brain Activity. US Patent Applicacion. US 2007/0191704 A1. 2007. [13] Bonfis P., Uziel A., Pujol R. “Screening for auditory dysfunction in infants by evoked otoacustic emission” Arch. Otolaryngol. Head Neck Surg, Vol. 114, 1988, pp. 887-90. [14] Gutierrez Gnecchi J.A., Herrera Garcia J. C. and Ortiz Alvarado J. D. “Auxiliary Neurofeedback System for Diagnostic of Attention Deficit Hyperactivity Disorder”. IEEE Conf. Electronics, Robotics and Automotive Mechanics Conference. 25-28 Sept. 2007, pp. 135-138. [15] Rompelman O. and Ros H.H. “Coherent averaging technique: A tutorial review. Part 1: Noise reduction and the equivalent filter. Part 2: Trigger Jitter, overlapping responses and nonperiodic stimulation”, J. Biomed. Eng., Vol. 8, 1986, pp. 24-35.

552