Elsevier

Clinical Neurophysiology

Volume 112, Issue 11, November 2001, Pages 1980-1998
Clinical Neurophysiology

Electrophysiological analysis of cortical mechanisms of selective attention to high and low spatial frequencies

https://doi.org/10.1016/S1388-2457(01)00660-5Get rights and content

Abstract

Objectives: This study investigated whether short-latency (<100 ms) event-related potential (ERP) components were modulated during attention to spatial frequency (SF) cues.

Methods: Sinusoidally modulated checkerboard stimuli having high (5 cycles per degree (cpd)) or low (0.8 cpd) SF content were presented in random order at intervals of 400–650 ms. Subjects attended to either the high or low SF stimuli, with the task of detecting targets of slightly higher or lower SF, respectively, than the above standards. ERPs were recorded from 42 scalp sites during task performance and spatio-temporal analyses were carried out on sensory-evoked and attention-related components.

Results: Attended high SF stimuli elicited an early negative difference potential (ND120) starting at about 100 ms, whereas attended low SF stimuli elicited a positivity (PD130) in the same latency range. The neural sources of both effects were estimated with dipole modeling to lie in dorsal, extrastriate occipital areas. Earlier evoked components evoked at 60–100 ms that were modeled with striate and extrastriate cortical sources were not affected by attention to SF. Starting at 150 ms, attended stimuli of both SFs elicited a broad selection negativity (SN) that was localized to ventral extrastriate visual cortex. The SN was larger over the left/right cerebral hemisphere for attended stimuli of high/low SF.

Conclusions: These results support the view that attention to SF does not involve a mechanism of amplitude modulation of early-evoked components prior to 100 ms. Attention to high and low SF information involves qualitatively different and hemispherically specialized neural processing operations.

Introduction

An important cue for visual selective attention is the size or spatial scale of a stimulus. For example, attention can be directed to emphasize perception of the parts (‘trees’) or the wholes (‘forest’) of a scene, both of which may occur in common regions of the visual field but at different spatial scales. One way to account for the attentional selection of parts and wholes is in terms of the selective enhancement of different spatial frequencies (SFs). There is abundant evidence suggesting that early in visual processing, retinal images are filtered by parallel SF selective channels (Cornsweet, 1970, De Valois and De Valois, 1990), and that attending selectively to gratings of a particular SF improves their detection and discrimination (Davis and Graham, 1981, Watson and Robson, 1981, Graham et al., 1985). Experiments using hierarchical stimuli comprised of a global figure made up of local elements have also provided evidence that high or low SFs (associated with the local and global levels, respectively) may be attended selectively (Shulman et al., 1986, Robertson, 1999). Accordingly, it has been proposed that the ability to selectively attend to high or low SFs is critical for visual object recognition (Ivry and Robertson, 1998).

The visual–cortical mechanisms underlying attention to SF have been studied by recordings of event-related potentials (ERPs). In the first such study, Harter and Previc (1978) recorded ERPs while presenting randomized sequences of checkerboards having 6 different check sizes as subjects attended to one check size at a time. Differences in the ERP waveforms elicited by attended vs. unattended stimuli were observed beginning at a latency of 160 ms in the form of an enlarged selection negativity (SN) to stimuli having the attended check size. Similar results were obtained in a series of studies in which gratings of high and low SF were presented in random order, with one of the frequencies being attended on any given run (Previc and Harter, 1982, Kenemans et al., 1993, Heslenfeld et al., 1997). Although ERPs to high and low SFs differed markedly in early-evoked components in the 60–140 ms range, selective attention did not reliably modulate these early components.

In the studies reviewed above, the earliest consistent effects of attention to SF cues were manifest as a broad SN over occipito-temporal scalp sites that onset in the 160–220 ms range accompanied by a frontal selection positivity (SP) beginning at 150–200 ms. Recently, however, Zani and Proverbio (1995) reported modulations of early, stimulus-evoked ERPs as a function of attention to SF. In their study, random sequences of 6 different checkerboards varying in SF (from 0.5 to 6 cycles per degree (cpd)) were presented. On different runs, subjects selectively attended to checkerboards of a single SF and ignored the remaining 5. The ERPs elicited by each checkerboard when it was task-relevant were compared to the averaged waveforms elicited by the same stimulus when it was task-irrelevant. Specifically, attended checkerboards yielded significantly enhanced amplitudes for a positive component peaking at 90 ms post-stimulus (termed P1 or P90). This component was largest at lateral occipital scalp sites and was also present in the unattended waveforms. Attended high SF checkerboards additionally elicited a significantly larger negativity with a medial occipital distribution peaking at 115 ms (N115). An SN with a peak latency of 270 ms was elicited by both high and low frequency patterns when attended.

The attention effects on these short-latency P90 and N115 components were interpreted by Zani and Proverbio (1995) as evidence that attention to SF involves an amplitude modulation of early-evoked cortical ERP components. They proposed that these attention effects reflected a modulation of sensory information flowing into SF channels and further that the enhanced N115 component might reflect attentional selection of SF information in the primary visual (striate) cortex, whereas the P90 modulation may originate in extrastriate visual cortex.

The findings of Zani and Proverbio (1995) stand in contrast with a number of previous studies, which concluded that early visual ERP components are only modulated during selective attention to location and not by attention to other stimulus features such as SF or color (reviewed in Mangun, 1995, Heslenfeld et al., 1997, Wijers et al., 1997, Hillyard and Anllo-Vento, 1998, Martinez et al., 2001). In these studies, stimuli at attended locations were found to elicit a characteristic ERP pattern that included an amplitude enhancement of the P1 (latency 80–120 ms) and N1 (140–200 ms) components relative to when the same stimuli were outside the ‘spotlight’ of spatial attention. Based on these findings, several authors have proposed that spatial attention acts by a sensory gain control mechanism reflected in P1/N1 amplitude modulations, which operates at an earlier stage of visual processing than does attention to non-spatial attributes (Hillyard and Munte, 1984, Heslenfeld et al., 1997, Wijers et al., 1997, Hillyard et al., 1998). Source localization studies have indicated that this enhancement of early P1/N1 activity during spatial attention takes place in extrastriate visual areas rather than in the primary visual cortex (Heinze et al., 1994, Clark and Hillyard, 1996, Gratton, 1997, Mangun et al., 1997, Woldorff et al., 1997, Martinez et al., 1999).

The findings reported by Zani and Proverbio, 1995, Zani and Proverbio, 1997 call into question the proposed uniqueness of spatial attention in exerting a gain control over sensory-evoked activity at the earliest levels of the visual pathways. To investigate this question further, the present experiment carried out multi-channel recordings and source localization of ERPs elicited during attention to sinusoidal gratings differing in SF. The task was designed to produce a state of highly focused attention to SF by presenting stimuli at a rapid rate and by requiring subjects to make a difficult discrimination of targets. The aim was to see whether early-evoked sensory activity is modulated by attention to SF in a manner resembling spatial attention effects.

A secondary aim of the present study was to obtain ERP evidence bearing on the hypothesis of hemispheric specialization for the processing of SF information (Christman et al., 1991, Kitterle et al., 1992). Behavioral studies using unilateral hemifield presentations have shown that when attention is selectively engaged in the right visual field (RVF) high SF stimuli are discriminated more readily than those of low SF content, and vice versa when attention is directed to the left visual field (LVF) (reviewed in Ivry and Robertson, 1998). Electrophysiological studies of attention to hierarchical letter stimuli have reported that selection of global targets is associated with a negative component (N2) having a greater amplitude over the right hemisphere (Heinze and Munte, 1993, Heinze et al., 1998); in contrast, the N2 was larger over the left hemisphere for local-level targets. The present study aimed to determine whether ERPs elicited during attention to sinusoidally modulated gratings of high and low SF are differentially distributed over the left and right hemispheres and to ascertain whether such asymmetries occur at early or late stages of processing (Ivry and Robertson, 1998).

Section snippets

Subjects

Twelve subjects (7 females, age range 19–33 years) participated as paid volunteers in the experiment. All subjects were right-handed as assessed by a brief questionnaire and had normal or corrected-to-normal vision. Each subject gave informed consent to participate in this experimental protocol, which was approved by the UCSD committee on human subject investigations.

Stimuli and task

Stimuli were sinusoidally modulated black and white checkerboard patterns, circular in overall form, presented on a

Behavioral performance

Responses to target stimuli of the attended SF made between 200 and 1200 ms following stimulus onset were considered ‘hits’. On an average, subjects correctly detected 85% of the high SF targets and 84% of the low SF targets. The mean RTs for detection of high and low SF targets were 533 and 545 ms, respectively. False alarms were calculated as the percentage of standard stimuli (of either the attended or unattended SF) that were followed by a response. The mean false alarm rate during the

Discussion

The present experiment investigated the spatio-temporal characteristics of the brain processes underlying selective attention to SF. The major focus was on finding out whether attention to SF produces a modulation of sensory-evoked neural activity beginning at the earliest stages of visual processing, as has been reported previously for attention to location. This issue was investigated by recording ERPs in a paradigm that required highly focused attention on the SF content of attended stimuli

Conclusions

In contrast with the reports of Zani and Proverbio, 1995, Zani and Proverbio, 1997 but in agreement with the studies of Heslenfeld et al., 1997, de Ruiter et al., 1998 and Kenemans et al. (in press), we found no evidence that attention to SF involves modulation of early-evoked activity in size-specific neural pathways in the 60–100 ms range. Instead, the earliest ERP signs of selection for SF began at around 100 ms in the form of increased negativity (ND120) to attended high SF stimuli and

Acknowledgements

This work was supported by NIMH grant MH-25594. We thank Matt Marlow for the technical assistance and 3 referees for their helpful comments on the manuscript.

References (51)

  • A. Martinez et al.

    Putting spatial attention on the map: timing and localization of stimulus selection processes in striate and extrastriate visual areas

    Vision Res

    (2001)
  • G. McCarthy et al.

    Scalp distributions of event-related potentials: an ambiguity associated with analysis of variance models

    Electroenceph clin Neurophysiol

    (1985)
  • M.B. de Ruiter et al.

    Effects of inter- and intramodal selective attention to non-spatial visual stimuli: an event-related potential analysis

    Biol Psychol

    (1998)
  • R.B. Tootell et al.

    From retinotopy to recognition: fMRI in human visual cortex

    Trends Cogn Sci

    (1998)
  • A.B. Watson et al.

    Discrimination at threshold: labelled detectors in human vision

    Vision Res

    (1981)
  • A. Zani et al.

    ERP signs of early selective attention effects to check size

    Electroenceph clin Neurophysiol

    (1995)
  • L. Anllo-Vento et al.

    Selective attention to the color and direction of moving stimuli: electrophysiological correlates of hierarchical feature selection

    Percept Psychophys

    (1996)
  • L. Anllo-Vento et al.

    Spatio-temporal dynamics of attention to color: evidence from human electrophysiology

    Hum Brain Mapp

    (1998)
  • V.P. Clark et al.

    Spatial selective attention affects early extrastriate but not striate components of the visual evoked potential

    J Cogn Neurosci

    (1996)
  • V.P. Clark et al.

    Identification of early visually evoked potential generators by retinotopic and topographic analysis

    Hum Brain Mapp

    (1995)
  • T.N. Cornsweet

    Visual perception

    (1970)
  • E. Davis et al.

    Spatial frequency uncertainty effects in the detection of sinusoidal gratings

    Vision Res

    (1981)
  • R.L. De Valois et al.

    Spatial vision

    (1990)
  • A.P. Ginsburg

    Spatial filtering and visual form perception

  • N. Graham et al.

    Attending to spatial frequency and spatial position of near-threshold visual patterns

  • Cited by (97)

    View all citing articles on Scopus
    View full text