Modulation of Gamma-Band Spectral Power by Cognitive Task Complexity Panagiotis G. Simos*, Elias Papanikolaou + , Evangelos Sakkalis + , and Sifis Micheloyannis + Summary: This study evaluated the utility of electroencephalographic (EEG) measures as indices of regional cerebral activation during performance of visuo-semantic analysis tasks in neurologically intact adult volunteers. EEG was recorded from 28 scalp locations as participants performed three visual discrimination tasks designed to tap into increasingly more complex operations regularly involved in the recognition of living animate objects. In addition, data from a control task involving the same stimuli, but requiring no cognitive decision or response, was included. EEG records were quantified using power spectrum measures in five frequency bands (delta, theta, alpha, beta, and gamma). Results showed a significant linear increase in absolute power in the gamma band with increasing task complexity over left hemisphere frontal and occipital regions, and over right temporoparietal regions. Key words: EEG; Power Spectrum; Gamma band; Cognition; Pictures; Visual analysis; Figural analysis. Introduction Electroencephalographic (EEG) measures have been used successfully in the past as indices of cerebral engage- ment in cognitive tasks. The majority of investigations have searched for systematic, task-related changes in the energy in the alpha band (8-13 Hz). There is evidence, however, that changes in spectral energy in higher (Krausse et al. 1998; Pulvermüller et al. 1997; Papanicolaou et al. 1986) frequency bands (such as beta -- 14 to 30 Hz -- and gamma bands - 30 to 60 Hz) are very sensitive as indi- ces of task-related cerebral activation. Briefly, EEG studies have shown alpha suppression (e.g., Makino 1986) and/or beta enhancement over the left hemisphere during verbal tasks (Bizas et al. 1999; Papanicolaou et al. 1986; Pulvermüller et al. 1997). Changes in gamma activity have been reported in a variety of per- ceptual/cognitive tasks (Keil et al. 1999; Pulvermüller et al. 1996; Krausse et al. 1998; Rodriguez et al. 1999). Specifically, activity in the gamma band has been reported in the context of face recognition tasks (Rodriguez et al. 1999), in tasks in- volving mental search of internal object representations (Tallon-Baudry et al. 1999), and in other complex visuo-perceptual tasks (Keil et al. 1999). The interest in non-invasive recordings of high-frequency electrical activ- ity in humans has grown considerably in recent years, after reports of similar high-frequency oscillations recorded invasively from large neuronal aggregates in animals (Gray and Singer 1989). The wide scalp distribution of task-related variations in gamma power and phase-locking raise the possibility that high frequency EEG activity re- flects synchronized oscillations in the human brain that form the basis of the engagement of neuronal populations in a particular function (Basar et al. 2000). In order to show, however, that variations in EEG spectral energy reflect engagement of neurophysiological processes specialized for particular cognitive operations, one must first demonstrate that power spectrum mea- sures vary systematically with the degree of complexity of the cognitive decisions required by the experimental task. Unfortunately, only a few studies have been de- signed along those lines (Bizas et al. 1999; Papanicolaou et al. 1986; Krausse et al. 1998). In the present study we explored the sensitivity of EEG-derived measures as indices of cerebral activation in areas involved in visuo-perceptual operations. In or- der to engage component operations that are critical for the decisions that require such operations, we manipu- lated the number of dimensions along which subjects were asked to make discriminatory decisions, while maintaining overall degree of visual complexity of the stimuli roughly constant across tasks. * Department of Neurosurgery, The University of Texas, Health Science Center, Houston, Texas, USA. + Clinical Neurophysiology Laboratory (L.Widen), Faculty of Medicine, University of Crete, Iraklion, Greece. Accepted for publication: October 12, 2001. Correspondence and reprint requests should be addressed to Dr. P.G. Simos, Department of Neurosurgery, The University of Texas Medical School, 6431 Fannin Suite 7.154, Houston, Texas, 77030, USA. Fax: (713) 500-0655 E-mail: psimos@uth.tmc.edu. Copyright  2002 Human Sciences Press, Inc. Brain Topography, Volume 14, Number 3, Spring 2002 (© 2002) 191