Alpha EEG coherence in different brain states: an electrophysiological index of the arousal level in human subjects Jose  L. Cantero a, b, * , Mercedes Atienza a, b , Rosa M. Salas a, b , Carlos M. Go  mez b a Laboratorio de Suen Ä o, A  rea de Psico®siologõÂa Cognitiva, Avenida de Andalucõ Âa 16, 18D-Izquierda, 41005 Seville, Spain b Laboratorio de Psico®siologõÂa, Departamento de Psicologõ Âa Experimental, Facultad de Psicologõ Âa, Universidad de Sevilla, Seville, Spain Received 26 May 1999; received in revised form 21 June 1999; accepted 30 June 1999 Abstract The functional relationships between the brain areas supposedly involved in the generation of the alpha activity were quanti®ed by means of INTRA- and INTER-hemispheric coherences during different arousal states (relaxed wakefulness, drowsiness at sleep onset, and rapid eye movement sleep) where such an activity can be clearly detectable in the human EEG. A signi®cant decrease in the fronto-occipital as well as in the inter-frontal coherence values in the alpha range was observed with the falling of the vigilance level, which suggests that the brain mechanisms underlying these coherences are state dependent. Making fronto-frontal coherence values in the alpha frequency band useful indexes to discern between brain functional states characterized by a different arousal level. q 1999 Elsevier Science Ireland Ltd. All rights reserved. Keywords: EEG coherence; Alpha activity; Wakefulness; Drowsiness; Rapid eye movement sleep; Alert index; Humans Alpha activity (7.5±12.5 Hz) is the electrophysiological correlate of a relaxed wakefulness state with closed eyes. Its brain generation sources have been proposed to be located in the occipital [10,16] and anterior regions [8,14]. This brain activity is also clearly observed in the human drowsi- ness period at sleep onset and in the rapid eye movement (REM) stage in a burst manner [3]. Few studies have explored the generation processes of human alpha activities appearing in other arousal states different from wakefulness. Hasan and Broughton [7] found that both the occipital and fronto-central areas played an independent role in the drow- siness-alpha generation. To our knowledge, there is no information available in the literature on the generation mechanisms of the REM-alpha bursts. However, this activ- ity has shown different properties of brain microstates [4] and seems to be functionally different from the alpha activ- ity contained in the REM-EEG background [5]. The functional relationships between different brain areas supposedly involved in the generation of an EEG activity can be studied by means of the coherence technique [1,6]. Results obtained with that technique and with correlation measures revealed that a functional interaction between anterior and posterior brain areas, physiologically sustained by the long fronto-occipital fasciculi [11], might be involved in the generation of the waking-alpha rhythm [18,20]. However, there is no quantitative information about relationships between those distant cortical areas in other arousal states where alpha activity is also clearly observable in the EEG. In a previous study, it was shown that human alpha activ- ity in relaxed wakefulness, drowsiness period, and REM sleep displayed a similar topographic distribution over the scalp [3]. In view of this ®nding, the aim of this work was to explore the possibility of the functional relationships between those brain areas supposedly involved in the generation of the alpha activity varying according to the brain activation state where such an activity appears spon- taneously. This may be suggested by changes in the associa- tion level between cortical regions, re¯ected by coherence values, depending on the arousal state. Ten normal subjects, ranging from 19±25 years, were examined. They slept in the sleep laboratory for 2 non- consecutive nights (®rst for adaptation). Polysomnography consisted of vertical and horizontal electrooculography (EOG), submental electromyography (EMG), and eight EEG channels (F3, F4, C3, C4, P3, P4, O1, and O2) refer- enced to linked mastoids. EOG and EEG bandpass were 0.5±40 Hz, whereas EMG ®lters were set in 5±70 Hz. Elec- trode impedance were always kept below 5 kV. In the morn- Neuroscience Letters 271 (1999) 167±170 0304-3940/99/$ - see front matter q 1999 Elsevier Science Ireland Ltd. All rights reserved. PII: S0304-3940(99)00565-0 * Corresponding author. Tel./fax: 134-95-4570-917. E-mail address: suesevilla@interbook.net (J.L. Cantero)