Evoked EEG alpha oscillations in the cat brain – a correlate of primary sensory processing?

Neuroscience Letters 240 (1998) 41–44

Evoked EEG alpha oscillations in the cat brain – a correlate of primary sensory processing? Martin Schu¨rmann a, Canan Bas¸ar-Eroglu b, Erol Bas¸ar a , c ,* a Institute of Physiology, Medical University Lu¨beck, 23538 Lu¨beck, Germany Institute of Psychology and Cognition Research, University of Bremen, 28334 Bremen, Germany c ¨ TUBITAK Brain Dynamics Research Unit, Ankara, Turkey

b

Received 10 November 1997; received in revised form 28 November 1997; accepted 1 December 1997

Abstract Event-related alpha (10 Hz) oscillations in the EEG were measured in cats by means of intracranial electrodes. Simultaneous recordings were made from auditory and visual cortex while auditory and visual stimuli were applied (in separate sessions). Frequency domain analysis of the EEG responses showed marked alpha components only for adequate stimulation (e.g. visual cortex-visual stimulus). This hints at a functional relationship between alpha responses and primary sensory processing.  1998 Elsevier Science Ireland Ltd.

Keywords: Alpha oscillations; Frequency domain analysis; Auditory cortex; Visual cortex

For a long time, alpha (10 Hz) rhythms (or oscillations) in the EEG have been regarded as correlates of ‘idling’. Recently, however, it has been suggested that alpha oscillations might have ‘closer relationships to ‘events’ than one might have thought earlier’ [15], possibly being related to ‘cortical work’ [13]. This view has been summarized in the expression ‘functional alpha’ [1,5]. Since stimulus-dependent 10 Hz oscillations have also been observed at the cellular level [7], alpha oscillations may soon attract as much interest as the widely discussed gamma oscillations. One of the functional correlates of alpha oscillations has become clear in cross-modality experiments, e.g. in visual cortex recordings of auditory and visual evoked potentials (EPs). The ‘alpha response’, obtained by digital filtering of the EP, was particularly dependent on whether the stimulus was adequate (in this case: visual) or not, suggesting a relationship between alpha response and primary sensory processing (in human scalp measurements [4] and in cat intracranial measurements [3]). In order to extend these results, we investigated frequency properties of EPs by means of a simple, yet effective quantitative evaluation of amplitude frequency characteristics [2]. * Corresponding author. Tel.: +49 451 5004170; fax: +49 451 5004171.

Measurements were performed in six cats. Electrodes were implanted, under Nembutal anesthesia, in several structures, among them the auditory cortex (gyrus ectosylvianus anterior, GEA) and the visual cortex (occipital cortex, OC; area 17) [6]. Three stainless steel screws in different regions of the skull served as a reference. Experiments started approximately two weeks after implantation, with cats freely moving or resting in a cage in a soundproof, dimly illuminated room. Auditory EPs (2000 Hz tones, 1 s duration, 0.5 ms rise time) and visual EPs (intense step-function stimuli delivered via a fluorescent bulb) were recorded (N = 100 stimuli each with pseudo-random intervals of 2.5–3.5 s in separate sessions). One-second pre- and post-stimulus EEGs (filter limits 0.5–70 Hz) were sampled at 500 points/s and digitized. EP frequency domain analysis [2], based on principles of systems analysis, aims at characterizing the ‘system’ brain by its input/output relation. The ‘classical’ approach to determine this relation is to stimulate the system with sinusoidal inputs of different frequencies f. The plot of output amplitude as a function of input frequency is called amplitude frequency characteristics (AFC), where large output amplitudes indicate resonance responses. Observation periods can be shortened when the AFC is computed via the Fourier transform G(jq) of the ‘impulse response’ l(t) eli-

0304-3940/98/$19.00  1998 Elsevier Science Ireland Ltd. All rights reserved PII S0304- 3940(97) 00926- 9

42

M. Schu¨rmann et al. / Neuroscience Letters 240 (1998) 41–44



n



∑ AFC(qk † =(n − m)

k =m

such that qm relates to f = 8 Hz (q = 2pf) and qn relates to f = 15 Hz (see [18] for an alternative procedure of quantitative AFC evaluation). Fig. 1 shows AFCs computed from auditory EPs. For GEA (left), consistent maxima in the alpha (8–15 Hz) frequency range with some inter-individual variation are observed. Correspondingly, the spectral grand average for GEA (bottom; computed by averaging individual AFCs in the frequency domain) shows a peak extending from approximately 7–15 Hz. In contrast, no consistent maxima are visible for OC recordings (right). No clearly dominant peaks are seen in the spectral grand average for OC (bottom). Fig. 2 shows AFCs computed from visual EPs. AFCs computed from GEA recordings (left) lack consistent maxima. Correspondingly, in the spectral grand average peaks in the 4–7 Hz range and in the 8–15 Hz co-exist, barely differing in amplitude. In contrast, high-amplitude peaks in the upper alpha range are visible in all but one of the AFCs for OC (right). In the spectral grand average, a maximum in the 8–16 Hz range is clearly dominant. A common feature of Figs. 1 and 2 may be put as follows: inadequate stimulation was related to AFCs without domiFig. 1. AFCs computed from auditory EPs recorded from the auditory cortex, GEA (left), and from the visual cortex, OC (right). AFCs obtained in 10 individual experiments and spectral grand average of these AFCs (bottom row) are shown. Along the abscissa: frequency in logarithmic scale; along the ordinate: amplitude, lG(j q)l, in decibels. The individual AFCs are normalized in such a way that the amplitude at 1 Hz is equal to 0 dB.

cited by a brief stimulus containing all possible input frequencies: G(jq) =

…∞

l(t)exp( − jqt)dt

0

where q = 2pf and AFC = lG(jq)l. The impulse response, in turn, is computed from the ‘step response’ c(t), i.e. from the EP, by means of differentiation: l(t) = d{c(t)}/dt. Numerically, an FFT algorithm is used to evaluate the N data points of a digitized time series Xn (Xn = X(Dt), n = 0…N − 1) of duration T = (N − 1)Dt as follows: Computation of Fourier coefficients: N −1

Yk = Y (qk ) = ∑ Xn exp( − j2pN − 1 nk); qk = 2pkT − 1 n=0

Computation of the AFC from the complex coefficients Yk = ak + jbk: AFC(qk ) = (a2k + b2k )1=2 For statistical evaluation, the values AFC(qk) were read and averaged within the frequency band of interest (e.g. 8–15 Hz range), giving the average amplitude

Fig. 2. AFCs computed from visual EPs recorded from the auditory cortex, GEA (left), and from the visual cortex, OC (right). AFCs obtained in 10 individual experiments and spectral grand average of these AFCs (bottom row) are shown. Axes and normalization as in Fig. 1.

M. Schu¨rmann et al. / Neuroscience Letters 240 (1998) 41–44

Fig. 3. Group averages of the average amplitude values read from the individual AFCs shown in Figs. 1 and 2 for the theta and alpha bands. Bars show the mean ± SD for the theta range (empty bars) and for the alpha range (hatched bars). Alpha responses of high amplitude are only observed with ‘adequate stimulation’ (manifested in significant theta-alpha differences).

nant maxima, whereas adequate stimulation led to AFCs with dominant maxima in the alpha range (alpha components). For quantitative evaluation, we computed average amplitudes for the alpha (8–15 Hz) range (from the individual AFCs) and compared them to average amplitudes for the theta range (4–7 Hz). The choice of range limits took into account the present grand average AFCs and results of our previous studies [4]. The results are summarized in Fig. 3. It was only for adequate stimuli (regardless of modality) that amplitudes for the alpha range were significantly larger than for the theta range (Wilcoxon tests of average amplitude for theta band versus average amplitude for alpha band). To sum up the results, dominant alpha components in the AFCs were only observed when the stimulus was adequate for the (primary sensory) area under study. This particular stimulus-dependence (distinguishing alpha components from other frequency ranges) hints at a functional relation between alpha components and primary sensory processing. Our results are consistent with previous studies in humans using EEG (high-amplitude alpha responses in occipital sites only with visual stimuli whereas theta responses were less stimulus-dependent [4]) and MEG [21]. (Note that alpha responses and alpha blocking may occur in the same subject depending on the amplitude of the pre-stimulus EEG [2,16] and that the spontaneous alpha rhythm in cats is less distinct than in humans.) Our results are related to the concept of event-related oscillations which has recently become popular. This may

43

be due to the fact that it served to link task-dependent gamma responses in the EEG to the widely discussed stimulus-dependent gamma oscillations at the cellular level [8,10]. For event-related alpha oscillations a similar development is under way [5]. Galambos [9] discerns the spontaneous alpha rhythm from several classes of event-related alpha oscillations with diverse functional correlates comprising memory and movement processes [12,17]. MEG studies permit a topographic differentiation of theta and alpha rhythms [11,14,19]. At the cellular level, stimulusdependent 10 Hz patterns of action potentials [7] show intriguing parallels to EEG oscillations. Alpha rhythms have also been related to feature integration over long cortical distances [20]. In summary, we used a frequency domain approach based on the working hypothesis of the EP as a superposition of event-related oscillations in several frequency ranges [2]. A supplementary procedure (confirming and extending results of digital filtering [4] and wavelet analysis [3]) served to demonstrate stimulus-dependent evoked alpha oscillations in primary sensory cortices. This hints at a functional relationship between alpha responses and primary sensory processing. ¨ BITAK Supported by DFG (Ba 831/5-1, 943/1-1) and TU (TBAG 17-1, 17-2, 17-3). [1] Bas¸ar, E., Brain Oscillations: Principles and Approaches, Springer, Berlin, 1997, in press. [2] Bas¸ar, E., EEG Brain Dynamics, Elsevier, Amsterdam, 1980. [3] Bas¸ar, E., Demiralp, T., Schu¨rmann, M., Bas¸ar-Eroglu, C. and Ademoglu, A., Oscillatory brain dynamics, wavelet analysis and cognition, Brain Cognit., 1997, in press. [4] Bas¸ar, E. and Schu¨rmann, M., Functional aspects of evoked alpha responses in humans and cats, Biol. Cybern., 72 (1994) 175–183. [5] Bas¸ar, E., Schu¨rmann, M., Bas¸ar-Eroglu, C. and Karakas, S., Alpha oscillations in brain functioning: an integrative theory, Int. J. Psychophysiol., 26 (1997) 5–29. [6] Bas¸ar-Eroglu, C., Bas¸ar, E. and Schmielau, F., P300 in freely moving cats with intracranial electrodes, Int. J. Neurosci., 60 (1991) 215–226. [7] Dinse, H.R., Kru¨ger, K., Akhavan, A.C., Spengler, F., Scho¨ner, G. and Schreiner, C.E., Low-frequency oscillations of visual, auditory and somatosensory cortical neurons evoked by sensory stimulation, Int. J. Psychophysiol., 26 (1997) 205– 227. [8] Eckhorn, R., Bauer, R., Jordan, W., Brosch, M., Kruse, W., Munk, M. and Reitboeck, H.J., Coherent oscillations: a mechanism of feature linking in the visual cortex?, Biol. Cybern., 60 (1988) 121–130. [9] Galambos, R., A comparison of certain gamma band (40-Hz) brain rhythms in cat and man. In E. Bas¸ar and T.H. Bullock (Eds.), Induced Rhythms in the Brain, Birkha¨user, Boston, MA, 1992, pp. 201–216. [10] Gray, C.M., Ko¨nig, P., Engel, A.K. and Singer, W., Oscillatory responses in cat visual cortex exhibit inter-columnar synchronization which reflects global stimulus properties, Nature, 338 (1989) 334–337. [11] Hari, R. and Salmelin, R., Human cortical oscillations: a neuro-

44

[12] [13] [14]

[15] [16]

[17]

[18]

M. Schu¨rmann et al. / Neuroscience Letters 240 (1998) 41–44 magnetic view through the skull, Trends Neurosci., 20 (1997) 44–49. Klimesch, W., Memory processes, brain oscillations and EEG synchronization, Int. J. Psychophysiol., 24 (1996) 61–100. Mulholland, T.B., Human EEG, behavioural stillness and biofeedback, Int. J. Psychophysiol., 19 (1995) 203–280. Narici, L., Pizzella, V., Romani, G. and Torrioli, G., Traversa, R. and Rossini, P.M., Evoked alpha and mu rhythm in humans: a neuromagnetic study, Brain Res., 520 (1990) 222–231. Niedermeyer, E., Alpha rhythms as physiological and abnormal phenomena, Int. J. Psychophysiol., 26 (1997) 31–49. Nogawa, T., Katayama, K., Tabata, Y., Ohshio, T. and Kawahara, T., Changes in amplitude of the EEG induced by a photic stimulus, Electroenceph. clin. Neurophysiol., 40 (1976) 78–88. Pfurtscheller, G., Event-related synchronization (ERS): an electrophysiological correlate of cortical areas at rest, Electroenceph. clin. Neurophysiol., 83 (1992) 62–69. Ro¨schke, J., Mann, K., Riemann, D., Frank, C. and Fell, J.,

Sequential analysis of the brain’s transfer properties during consecutive REM periods, Electroenceph. clin. Neurophysiol., 96 (1995) 390–397. [19] Saermark, K., Mikkelsen, K.B. and Bas¸ar, E., Magnetoencephalographic induced rhythms. In E. Bas¸ar and T.H. Bullock (Eds.), Induced Rhythms in the Brain, Birkha¨user, Boston, 1992, pp. 129–145. [20] Schanze, T. and Eckhorn, R., Phase correlation among rhythms present at different frequencies: spectral methods, application to microelectrode recordings from visual cortex and functional implications, Int. J. Psychophysiol., 26 (1997) 171–189. [21] Schu¨rmann, M., Bas¸ar-Eroglu, C., Rahn, E., Braasch, M., Do¨ssel, O., Fuchs, M. and Bas¸ar, E, A comparative study of alpha responses in human MEG temporo-parietal and occipital recordings and cat intracranial recordings, Proceedings of the IEEE Satellite Symposium on Neuroscience and Technology, Lyon, 1992, pp. 132–137.