Elsevier

Brain Research

Volume 1172, 3 October 2007, Pages 124-129
Brain Research

Research Report
Welcome to the real world: Validating fixation-related brain potentials for ecologically valid settings

https://doi.org/10.1016/j.brainres.2007.07.025Get rights and content

Abstract

Exploration of the real world usually expresses itself through a perceptual behaviour that is complex and adaptive — an interplay between external visual and internal cognitive states. However, up to now, the measurement of electrophysiological correlates of cognitive processes has been limited to situations, in which the experimental setting confined visual exploration to the mere reception of a strict serial order of events. Here we show – exemplified by the well known old/new effect in the domain of visual word recognition – that an alternative approach that utilizes brain potentials corresponding to eye fixations during free exploration reveals effects as reliable as conventional event-related brain potentials.

Introduction

The most prominent advantage of event-related potentials (ERPs) is that the time course of cognitive processes can be measured at a very high temporal resolution (Hillyard and Kutas, 1983). A critical cornerstone for the analysis of brain potentials is determining the onset of the theoretically relevant cognitive process the electrophysiological correlates of which are supposed to be recorded. In conventional ERP paradigms, a cognitive process and the recording of its electrophysiological correlate are synchronized by linking them to a single external event — most often an experimentally presented stimulus. This approach is based on the implicit assumption that a specific cognitive process (e.g., the recognition of a word) can not start until the relevant information (e.g., the word) – on which the specific cognitive process is dependent – is provided. By having knowledge about the point in time at which the relevant bit of information was provided researchers know where in the recorded data-stream to look for the electrophysiological correlates of the cognitive process of interest. Although this approach is suitable for the synchronization of cognitive processes with the recording of electrophysiological data, it also bears important limitations with respect to the ecological validity of the resulting experimental settings: Experience and behaviour is broken down into a serial sequence of externally triggered events. For example, in the domain of research on reading and dyslexia, the presentation of a sentence is segmented into an artificial sequence of isolated words, with each word being presented individually, one after the other, separated by relatively long time intervals (usually about 500 ms). Although being useful for research on single word recognition, the experimental paradigms used within the ERP framework are far from natural reading of meaningful text.

Here we present the validation of an alternative approach that allows synchronizing cognitive processes and the recording of its electrophysiological correlates by relating a natural sequence of perceptual events to a series of eye movements and fixations: the fixation-related potentials (FRP) approach. Contrary to the ERP paradigm, in the FRP approach synchronization is not realized by the externally triggered presentation of a stimulus. Rather, an eye tracker is used to assess a subject's eye movements in order to determine when a subject is looking at what (e.g., a word) in a complex, visual pattern (e.g., a sentence).

By doing so, the exact point in time at which certain visual information (that is crucial for triggering a specific cognitive process) is taken in can be assessed — allowing to synchronize the recording of electrophysiological data to the start of a specific cognitive process. Monitoring a person's eye movements during the recording of electrophysiological data therefore provides a self-paced but externally observable indicator for the beginning of a cognitive process (that relied upon the perception of that very specific visual matter).

For a wide applicability of the FRP technique, the exact determination of a subject's gaze position by means of an eye tracker is inevitable. Prior to the broad availability of eye trackers, electro-oculographic (EOG) recordings were used to assess the on- and offset as well as the direction of saccades. Utilizing EOGs, Marton et al. (1985) could show that brain potentials corresponding to saccades were comparable to visual-evoked potentials in their late components. However, because determining a subject's absolute gaze position by means of EOG is difficult, experimental implementations of this approach bare severe restrictions. For example, during sentence reading only brain potentials corresponding to the last word in a sentence could be assessed (e.g., Marton and Szirtes, 1986, Marton and Szirtes, 1988a, Marton and Szirtes, 1988b).

With the rise of easy-to-use eye trackers it is now unproblematic to determine a subject's gaze position with high spatial resolution while recording electrophysiological data. Although already utilized for the assessment of cognitive processes (Baccino and Manunta, 2005, Dimigen et al., 2006) a validation of the FRP is necessary in order to prove that the electrophysiological correlates of cognitive processes as reflected in conventional ERPs are also evident in FRPs.

The aim of the present paper is to validate the FRP approach. For that purpose, the underlying rationale was to cross validate an effect that is well documented in conventional ERP research: the old/new effect in visual word recognition, characterized by a positivation for correctly recognized old words from around 250 ms (Rugg and Nagy, 1989, Rugg et al., 2000). This way we can directly compare findings from FRPs with the ERP marker, which in turn is established by state-of-the-art literature. Specifically, a continuous recognition task (Friedman, 1990) was realized in both, the conventional ERP paradigm (with five words being presented one at a time) and in the FRP paradigm, in which the same participants were presented with five words all at once in a row while their eye movements were recorded. Finding the old/new effect in FRPs with a spatial distribution comparable to that found in the ERPs would argue for the validity of the FRP approach.

Section snippets

Analysis

For ERPs, continuous EEG data was segmented according to the standard procedure: Upon the point in time of appearance of the theoretically relevant target word, segments from 100 ms pre-stimulus to 600 ms post-stimulus were extracted for analysis. For segmentation of EEG data in the FRP setting, the point in time of a participant's first fixation on the target word was determined. Based upon the synchronization of the EEG and eye-movement recording, the onset time of the respective fixation

Discussion

The aim of the present paper was the validation of the FRP approach. The important finding is that the chosen marker effect – i.e., the old/new effect – was observed in FRPs as well as in ERPs. Moreover, the spatial distribution of the old/new effect was the same in FRPs and ERPs: In both paradigms, the magnitude of the old/new effect was greater in the right than in the left hemisphere (in accordance with evidence for shallow encoding reported by Rugg et al., 2000). Furthermore, for FRPs and

Subjects

Twenty students (12 female) of the Freie Universität Berlin (mean age 24;6 [years;month]) participated in the study, all were native German speakers with normal or corrected to normal vision, and all except one were right handed.

Experimental procedure

To obtain the electrophysiological correlates of the old/new effect, a continuous recognition paradigm (Friedman, 1990) was applied: A series of five words was presented and participants had to indicate via button press, whether the fifth word – the target word – was

References (14)

There are more references available in the full text version of this article.

Cited by (79)

  • Lateralization of early orthographic processing during natural reading is impaired in developmental dyslexia

    2022, NeuroImage
    Citation Excerpt :

    In contrast to fixed-gaze reading when the order in which participants have to process the words is controlled by the experimenter, in natural reading paradigms whole text lines or paragraphs are presented and participants are allowed to scan the words in an arbitrary order that could also include refixations. The deflection of occipito-temporal FREA after the first positive peak (P1 or lambda component) is shown to be shifted into the positive direction in the case of refixations compared to first fixations during natural reading (Loberg et al., 2019) as well as during fixations on identical word stimuli in saccadic reading of unconnected words (Dimigen et al., 2012; Hutzler et al., 2007). As these effects might exhibit left-hemispheric dominance around the first negative FREA component (Dimigen et al., 2012), a reduction of left-hemispheric FREA lateralization could be predicted at this latency during refixations in control readers.

View all citing articles on Scopus
View full text