Brain electric microstates and momentary conscious mind states as building blocks of spontaneous thinking: I. Visual imagery and abstract thoughts

Abstract

Prompted reports of recall of spontaneous, conscious experiences were collected in a no-input, no-task, no-response paradigm (30 random prompts to each of 13 healthy volunteers). The mentation reports were classified into visual imagery and abstract thought. Spontaneous 19-channel brain electric activity (EEG) was continuously recorded, viewed as series of momentary spatial distributions (maps) of the brain electric field and segmented into microstates, i.e. into time segments characterized by quasi-stable landscapes of potential distribution maps which showed varying durations in the sub-second range. Microstate segmentation used a data-driven strategy. Different microstates, i.e. different brain electric landscapes must have been generated by activity of different neural assemblies and therefore are hypothesized to constitute different functions. The two types of reported experiences were associated with significantly different microstates (mean duration 121 ms) immediately preceding the prompts; these microstates showed, across subjects, for abstract thought (compared to visual imagery) a shift of the electric gravity center to the left and a clockwise rotation of the field axis. Contrariwise, the microstates 2 s before the prompt did not differ between the two types of experiences. The results support the hypothesis that different microstates of the brain as recognized in its electric field implement different conscious, reportable mind states, i.e. different classes (types) of thoughts (mentations); thus, the microstates might be candidates for the `atoms of thought'.

Introduction

Brain electric field recording allows the monitoring of brain work continually and non-invasively with that very high time resolution which is needed to study cognitive–emotional processes. At each time instant, the data yield a map of the potential distribution on the head surface (Lehmann, 1971). Thus, brain activity can be visualized as a series of momentary brain electric field maps. An instantaneous map reflects the sum of all momentarily active brain processes, superficial and deep (Smith et al., 1983). If the map's spatial configuration (the momentary landscape) changes, different neural elements must have become active. It appears reasonable to assume that different sets of active neural elements perform different functions.

Using data-driven analysis strategies we found that the changes of the spatial configuration of the brain field are discontinuous; they occur in a step-wise fashion. Accordingly, the continuous stream of maps of momentary electric field potential distributions can be segmented into time epochs of varying durations in the sub-second range during which the field shows a near-stable landscape (Lehmann and Skrandies, 1980, Lehmann, 1984). These epochs of quasi-stable field landscape were called microstates and were observed with very different analysis approaches and experimental conditions during event-related brain activity (Lehmann and Skrandies, 1980, Brandeis and Lehmann, 1989, Michel et al., 1992, Brandeis et al., 1995, Pascual-Marqui et al., 1995, Koenig and Lehmann, 1996, Fallgatter et al., 1997, Kondakor et al., 1997Pegna et al., in press) as well as during spontaneous brain activity (Lehmann, 1984, Lehmann et al., 1987, Merrin et al., 1990, Strik and Lehmann, 1993, Wackermann et al., 1993, Koukkou et al., 1994, Kinoshita et al., 1995).

It thus appears that continuous brain electric activity consists of a concatenation of building blocks, the microstates, which are defined by their quasi-stable field landscapes and which are suggested to incorporate different modes, contents or steps of information processing. This raises the suggestion that the subjective experience of what William James called the `stream of consciousness' actually consists of discernible elements. Of course it is clear that this stream of consciousness over time carries varying contents as has been studied repeatedly (see Pope and Singer, 1978). During free-floating, no-task conditions (so-called day dreaming), experiences of visual imagery are frequent and reports of subjective experiences involving visual imagery are easy to distinguish from reports of non-imaginal thoughts which concern abstract topics such as planning (Foulkes and Fleisher, 1975, Lehmann et al., 1995). The PET cerebral correlate of visual imagery was also reported to differ from that of non-imaginal thinking (Goldenberg et al., 1989).

The brain mechanisms of visual imagery were the topic of many studies (see Pylyshyn, 1981, Paivio, 1986, Kunzendorf and Sheikh, 1990, Kosslyn, 1994). Early general hypotheses on right hemispheric specialization for visual imagery, typically based on EEG power measurements (for a very critical overview see Ehrlichman and Barrett, 1983) were replaced by proposals that different modes of visual imagery utilize different brain mechanisms involving not solely the occipital regions (Roland and Friberg, 1985, Petsche et al., 1992) and involved left- as well as right-predominant mechanisms. Input-driven imagery tasks were reported to cause left-sided occipital event-related potential maxima (Farah et al., 1989), similar to PET studies where task solution or attention to images caused stronger left-sided activity, but instruction-less or memory-based imagery was more right-sided (e.g. Goldenberg et al., 1987, Kosslyn et al., 1995). Mental rotation, a particularly well studied imagery operation, typically showed a right-hemisphere preponderance (e.g. Papanicolaou et al., 1987, Corballis and Sergent, 1989Pegna et al., in press), similar to other examined imagery conditions (e.g. Sergent, 1989) and to task-free visual input event-related potentials (Mecacci et al., 1990).

Investigations of the microstate structure of event-related potential map series lead to the identification of certain microstates with certain processing steps (e.g. by Brandeis and Lehmann, 1989, subjective visual contours; Brandeis et al., 1995, congruent vs. incongruent sentence endings; Koenig and Lehmann, 1995, reading of abstract vs. imagery words; Koenig and Lehmann, 1996, reading of verbs vs. nouns; Pegna et al., in press, mental rotation).

The topic of the present article is the functional significance of different types of brain electric microstates during spontaneous, `free-running' brain activity. A no-task, no-input, no-response paradigm was chosen in order to avoid overlays of potentially influential brain subroutines that are necessary for continual remembering of and attending to a task, task execution and preparation and implementation of motor acts (Antrobus, 1987) and in order to examine mind states as such without externally driven representations of information. Hence, the article examines the concept that spontaneous, conscious states of the mind experienced as recallable thoughts can be described as physical states of the brain.

The present study presents evidence suggesting that two specific, spontaneous, conscious mind states are associated with two different classes of brain microstates that are defined in the brain electric field. In our no-input, no-task, no-response paradigm, subjects (when hearing a random prompt) reported recall of their immediately preceding spontaneous, conscious, unconstrained, private experiences. The reports were classified into experiences involving visual imagery or abstract thinking. An earlier analysis which used the entire 2-s epochs before the prompts showed that the two classes of subjective reports were associated with different locations of frequency domain EEG source models (Lehmann et al., 1993). The present analysis shows that the two mentation classes were associated with two different classes of brief brain electric field microstates immediately before the prompt signal, but not 2 s earlier.