In two EEG experiments using two rows of electrodes, Jeffreys and Axford (1972a, 1972b) described the characteristics of the earliest components of the VEP in humans.
As opposed to most studies before and since, they presented stimuli peripherally, rather than foveally or on the visual field meridians. In addition, they systematically tested for differences between stimuli restricted to different quadrants and octants, with a particular focus on variations between upper and lower visual fields. Based on their results (see also Jeffreys, 1971) — including some of the earliest dipole simulations of human EEG data – and standard descriptions of the anatomy of the human visual system (Holmes, 1945), they concluded that the first of these components, characterised by an onset around 50 ms post-stimulus and peak latencies substantially below 100 ms, originated in V1.
Although it has since been demonstrated that higher, extrastriate visual cortex is already active at the peak latency of this earliest VEP component (Foxe & Simpson, 2002), the basic assertion that what Jeffreys and Axford (1972a) termed “C. I” (for
”component I”) has its principal sources in V1 has stood the test of time (Clark, Fan,
& Hillyard, 1995; Di Russo, Martinez, Sereno, Pitzalis, & Hillyard, 2002; Im, Gurura-jan, Zhang, Chen, & He, 2007; Martinez et al., 1999; Pourtois, Grandjean, Sander, &
Vuilleumier, 2004). The basic model proposed by Jeffreys and Axford (1972a) convinc-ingly explains the most striking feature of the C1, which may also be the reason why the provisional name stuck: the component reverses its polarity depending on whether the upper or lower visual field is stimulated (see Fig. 1.1). Consequently, the stan-dard nomenclature of numbered positive- and negative-going peaks does not fit. Some researchers have labeled the component as the NP80 (for ”negative-positive”; Lange, Wijers, Mulder, & Mulder, 1998; Wijers, Lange, Mulder, & Mulder, 1997) to indicate the distinctive polarity reversal. According to the Jeffreys and Axford model, this po-larity reversal is due to the fact that V1 in the human principally covers the upper and lower banks of the calcarine sulcus, with the inversion between the external visual field and its central representation (Holmes, 1945) dictating that the upper visual field
Introduction
is represented on the lower bank of the calcarine and the lower visual field on the up-per bank. Assuming a “canonical” calcarine running orthogonal to the interhemispheric fissure and roughly parallel to the parieto-occipital scalp, selective stimulation of the upper or lower visual field should excite populations of cortical neurons with opposing orientations, leading to the observed polarity reversal. Upper visual field stimulation then elicits a surface-negative component, and lower visual field stimulation leads to a surface-positive potential.
Figure 1.1: Recording setup and VEPs for simple upper and lower visual field stimulation as used by Jeffreys and Axford (1972a). Note the polarity inversion of the first component, which the authors termed
”C. I”. Positive voltages are up.
This model, also explains why a C1 component is absent in most VEP studies:
if stimuli are presented foveally, their canonical representation should predominantly fall onto the outer occipital surface. As the representation of the fovea shows large differences in terms of size and location between individuals (Dougherty et al., 2003), any VEP components at this very early latency are effectively smoothed out. On the other hand, many authors presented stimuli centered on the horizontal meridian, whose representation usually falls into the fundus of the calcarine sulcus. In the case of bilateral stimulation, this would create opposing electrical fields in the two hemispheres that may cancel each other at the level of the scalp. In the case of unilateral stimulation, one may expect a bilaterally distributed component, but this has rarely been reported (however, see Im et al., 2007, below).
Several limitations need to be noted concerning this model, and Jeffreys and Axford were well aware of them, describing it as “almost certainly an oversimplification for most subjects” (Jeffreys & Axford, 1972a, p.18). First and foremost, individual differences in functional visual cortex anatomy are not limited to the representation of the fovea, but equally concern the location and extent of V1 and extrastriate visual areas, as well as the shape of the calcarine sulcus itself (Amunts, Malikovic, Mohlberg, Schormann, & Zilles,
Introduction
2000; Dougherty et al., 2003; Hasnain, Fox, & Woldorff, 1998). Accordingly, atypical C1 topographies are frequently observed (Fig. 1.2), making C1 measurements notoriously difficult to compare across subjects and studies (Proverbio, Del Zotto, & Zani, 2007).
A second, related caveat concerns the representation of the horizontal meridian, which in the majority of subjects does not seem to coincide with the fundus of the cal-carine, as its selective stimulation often evokes a negative C1 (Aine et al., 1996; Clark et al., 1995). This is in accordance with known differences between the cortical repre-sentations of the upper and lower visual field in humans: higher contrast sensitivity and spatial resolution are observed in the lower visual field (Lehmann & Skrandies, 1979;
Liu, Heeger, & Carrasco, 2006; Skrandies, 1987; Talgar & Carrasco, 2002), correspond-ing to a larger extent of the neuronal populations representcorrespond-ing this part of the visual environment in primates (Van Essen, Newsome, & Maunsell, 1984). It has been argued that environmental factors are responsible for this relative over-representation of the lower visual field (Skrandies, 1987). For example, higher spatial frequencies and lower contrast are usually present in the lower visual field of humans as we move through the environment and our visual systems may thus be hard-wired or trained to process such information more efficiently by expanding the neural populations coding for the lower visual field. The representation of the latter would thus encroach onto the lower back of the calcarine, producing the frequently observed negative C1 following stimulation on the horizontal meridian.
Figure 1.2: Single subject (left) and grand average (right) C1 from Experiment 1. Data were acquired following stimulation in the lower visual field. The subject shown on the left was excluded from analysis due to the atypical topography of the C1. Topographies are shown for the highlighted intervals; the time-window for the grand-average C1 is slightly larger due to blurring of the component when calculated across subjects, but equivalent differences in topography were observed for single time-frames. For details, see Results.
Notwithstanding these limitations, the general model as proposed by Jeffreys and Axford (1972a) is still frequently cited to explain the characteristics of the C1 (Clark et al., 1995; Di Russo et al., 2002; Martinez et al., 1999; Pourtois et al., 2004). Following the original report (Jeffreys & Axford, 1972a), a number of studies have explicitly addressed the component’s characteristics and their results have led to the predominant view that C1 amplitude and latency are exclusively a function of physical stimulus characteristics
Introduction
and not subject to top-down influences. Before providing an overview of these studies in Section 1.3, I will briefly address an important conceptual issue.