Brain, Vol. 124, No. 7, 1417-1425,
July 2001
© 2001 Oxford University Press
Contextual guidance of attention
Human intracranial event-related potential evidence for feedback modulation in anatomically early temporally late stages of visual processing
1 Department of Psychology, Yale University, 2 Department of Neurology, Yale University School of Medicine, New Haven, Connecticut, 3 Department of Psychology and Vision Research Center, Vanderbilt University, Nashville, Tennessee and 4 Neuropsychology Laboratory, Veterans Administration Medical Center, West Haven, Connecticut, USA
Correspondence to:
Ingrid R. Olson, Department of Diagnostic Radiology, Yale University School of Medicine, 333 Cedar Street, PO Box 208042, New Haven, CT 06520-8042, USA E-mail: iolson{at}boreas.med.yale.edu
 |
Abstract |
|---|
 Top Abstract IntroductionMethodResultsDiscussionAcknowledgementsReferences |
|---|
We investigated attentional guidance in early visual areas in the brain by recording event-related potentials directly from the surface of visual cortex. Patients performed a contextual cueing task in which attentive search to targets was guided by implicitly learned spatial context information. The earliest activity in striate cortex (area V1) was not modulated by contextual cueing, whereas later activity beginning at ~200 ms was enhanced by contextual cueing in V1, V2 and other portions of extrastriate cortex. These results suggest that context can enhance visual processing by temporally late top-down modulation of activity in anatomically early areas of visual cortex. Together with anatomical and neurophysiological studies in animals, these results suggest an excitatory feedback mechanism acting on apical dendrites of pyramidal cells in V1 and other areas of visual cortex.
event-related potential; attention; context; top-down modulation; visual cortex
AUC = area under the curve; ERP = event-related potential; FG = fusiform gyrus; IOG = inferior occipital gyrus; ITG = inferior temporal gyrus; V1 = visual cortex
|
Introduction |
|---|
|
TopAbstractIntroductionMethodResultsDiscussionAcknowledgementsReferences |
|---|
How does the visual system determine where to look? This has been a central question of visual attention research, and various factors have been found to facilitate detection of a target amongst distractors in visual search tasks. Bottom-up influences that guide the deployment of attention are salient features (Bravo and Nakayama, 1992
), abrupt onsets (Yantis and Jonides, 1984) and singleton features (Treisman and Gormican, 1988). Top-down, or knowledge-based factors that can guide attention, are novelty (Johnston et al., 1990), familiarity (Wang et al., 1994) and expectancy (Shaw and Shaw, 1977; Shaw, 1978; Miller, 1988). Here, we focus on how top-down contextual information directs attention. Recently, it has been shown that learned spatial context information can guide attention to important locations (Chun and Jiang, 1998; Chun, 2000). Contextual guidance of attention facilitates detection of targets amongst distractors in visual search tasks, reflecting top-down processing based on implicit memory traces of past views. How do memory-based representations guide attention and search? Does modulation occur at an early stage of visual processing or later?
Early or late modulation by attention?
Researchers have investigated the first point of attentional modulation in the visual processing pathway. Many different studies found that spatial attention modulates extrastriate areas but not striate cortex (V1), using a variety of techniques such as single-cell recording in monkeys (Luck et al., 1997), recording blood flow or electrical activity in humans (Heinze, 1994; Mangun et al., 1997; Hillyard and Anllo-Vento, 1998; Kastner et al., 1998; Mangun et al., 1998), or optical imaging in humans (Gratton, 1997).
However, more recent studies suggest that spatial attention may modulate V1 activity (Roelfsema et al., 1998; Vidyasagar, 1998; Watanabe et al., 1998; Gandhi et al., 1999; Ito and Gilbert, 1999; Martinez et al., 1999; Somers et al., 1999). Interestingly, some of these studies suggest that attentional modulation occurs by a temporally late mechanism. For instance, attention affects processing in macaque V1 at 100–200 ms (Vidyasagar, 1998), later than the initial sensory processing in macaque V1 at 34–94 ms (Schmolesky et al., 1998). A similar conclusion was reached using both functional MRI and scalp event-related potentials (ERPs) in humans (Aine et al., 1995; Martinez et al., 1999). These findings suggest an important point: attentional modulation may occur in an anatomically early area by a temporally late mechanism.
Temporally late modulation in an anatomically early area can be taken as evidence for re-entrant processing. Top-down attentional guidance may be one common cognitive operation that uses this mechanism (see also Zipser et al., 1996). We investigated this possibility in humans using intracranial ERPs, which allow recording directly from the cortical surface of the electrical events associated with a perceptual or cognitive process. Although scalp ERPs provide valuable details about the time course of perceptual and cognitive processing, this technique has poor spatial resolution. Functional MRI has excellent spatial resolution but poor temporal resolution and is an indirect measure of neural changes. Recording ERPs from the cortical surface and deep structures of patients with intractable epilepsy provides the spatiotemporal resolution necessary to make inferences about the neural events involved in perceptual and cognitive processes (Halgren et al., 1980; Nobre et al., 1994) and the temporal resolution and sensitivity needed to make statements about processing level and order (Allison et al., 1999). Here we summarize the results of the first intracranial ERP study of top-down contextual modulation of neural activity in human visual cortex.
Context and attention
We used the contextual cueing paradigm to study top-down, memory-based attentional guidance (Chun and Jiang, 1998). Subjects were required to look for a target in a visual search display (Fig. 1). Visual search is a commonly used spatial attention manipulation. Previous investigators have used visual search to study spatial attention as measured by scalp ERPs (Luck and Hillyard, 1994a, b). In contextual cueing, the distractor items, or the `context', are arranged to form spatial configurations that are repeated over time (Old condition). Subjects find targets more quickly when their location is associated with a repeated spatial context, compared with targets located in novel contexts (New condition), even though they are unaware of the repeated context. Thus, contextual cueing refers to the guidance of attention by implicit memory traces of context information.
|
The mechanisms of contextual cueing are unknown. Although an intact hippocampus and medial temporal lobe system appears to be important for storing spatial context information (Chun and Phelps, 1999
), the process of memory-based, top-down attentional deployment is likely to be performed by other areas in the brain. The question is whether such contextual guidance of attention occurs in early visual areas such as V1 and V2 or in later stages of visual processing. Although search tasks are not as direct as cueing tasks for measuring the deployment of attention, Luck and Hillyard have established that scalp ERP signs of attentional enhancement to targets are the same in search and cueing tasks (Luck and Hillyard, 1994a). Hence, by measuring both the timing and location of intracranial ERPs, inferences can be made about the mechanism of contextual modulation of attention. First, we examine whether memory-based context information influences activity in V1 and V2. Secondly, if modulation is observed, what is its time course? If contextual guidance modulates neural activity at early stages of cortical processing, a temporally early potential, peaking at ~100 ms in V1 and V2 (Allison et al., 1999), should be affected. However, attentional guidance that modulates activity after 100 ms in these areas would suggest that top-down processing operates by re-entrant feedback from higher cortical areas to early visual areas. |
Method |
|---|
|
TopAbstractIntroductionMethodResultsDiscussionAcknowledgementsReferences |
|---|
Subjects
Eighteen patients (15 right-handed, average age 29 years) with medically intractable epilepsy, who were being evaluated for possible resection surgery (Spencer et al., 1982
), served as subjects. All were patients of the Yale-New Haven Hospital Epilepsy Surgery Program. Prior to resection surgery, electrodes were implanted to localize the seizure focus. The protocol used was approved by the Human Investigations Committee of the Yale University School of Medicine. Informed consent was obtained. Subjects volunteered to participate and understood the experimental nature of the task. Five patients who did not show a contextual cueing benefit and one patient who had severe EEG artefacts were excluded from the analysis.
Electrode placement and localization
Under general anaesthesia, strips or grids of electrodes were placed subdurally on the lateral and ventral surface of the brain. The exposed surface of each stainless steel electrode was 2.2 mm in diameter; inter-electrode spacing was 10 mm. Electrodes were maximally sensitive to field potentials generated within the 3.8 mm2 area in contact with the brain, but were less sensitive to activity generated in surface cortex more than a few millimetres distant from the electrode. Depth probes were inserted from the lateral temporal lobe and were targeted to the hippocampus and other medial temporal lobe structures (McCarthy et al., 1991); each electrode was a cylinder 1 mm in diameter, 2 mm long, spaced 9 mm apart. Each patient had electrodes placed on portions of the occipital, temporal, parietal and frontal lobes. Procedures used to determine electrode locations are described in detail elsewhere (Allison et al., 1999). Briefly, this was performed by localizing electrodes in T1-weighted three-dimensional MRIs obtained the day following implantation. Locations were translated into the standardized coordinate system of Talairach and Tournoux (Talairach and Tournoux, 1988). Standardized Talairach space was used to characterize electrode locations along the y-axis. However, the use of Talairach space in the x-axis would lead to incorrect anatomical characterization due to individual anatomical variation in the configuration of gyri and sulci. Thus in the medial–lateral dimension, we used major landmarks for localization. Coronal images allowed identification of electrode sites in V1 (cuneate and superior lingual gyri), V2 (inferior lingual gyrus) and extrastriate areas (fusiform, inferior occipital and inferior temporal gyri).
ERPs
Recordings were made simultaneously from 64 electrodes per patient. ERP recordings were obtained concurrently with clinical EEG and behavioural video recordings, and were time-locked to stimulus onset. ERP recordings were not obtained immediately before or after seizures, and the EEG was normal at recording sites of interest. An experimenter sat at the patient's bedside to monitor general fixation and alertness. Patients were free to scan the display. Possible EEG contamination by the EOG is not a problem in intracranial compared with scalp recordings due to the order of magnitude larger ERPs recorded intracranially. Bandpass was 0.1–100 Hz (–3 dB points), gain was 10 000 and digitization rate was 250 Hz. Recordings were referential to a surface electrode at the mastoid, and were obtained 3–7 days post-implant. ERP waveforms for each condition and set were stored for later averaging.
Task
The experiment began with instructions followed by a practice block of 12 trials to familiarize patients with the task. They were not informed that the spatial configurations of the stimuli would be repeated on some trials, nor were they told to attend to the global array. They were simply instructed to search for a rotated `T' target among the rotated `L' distractors (Fig. 1
). A target was present in every trial. Patients were instructed to respond as quickly and as accurately as possible. Each trial began with a small fixation cross in the centre of the screen. The display appeared 200 ms later. After finding the target, they indicated the direction of the target, either left or right, by pressing a corresponding button, which cleared the screen for the next trial. No feedback was given. A mandatory break of 8 s was given at the end of each 32 trials. At the end of the break, the task resumed by itself.
The search display contained an equal number of red, blue, green and yellow items, 12 in total, which appeared within an invisible grid of 8 x 6 locations (Fig. 1). The background was grey. The base of the `T' target pointed either left or right. The L shapes were presented randomly in one of four orientations (0, 90, 180 or 270°. The visual search array subtended ~37 x 28° of visual angle. The size of the stimuli was ~2 x 2°. The position of each item was jittered within the rectangular array to prevent colinearities with other stimuli. The jittered position for each item was held constant throughout the experiment for Old arrays. Patients viewed the stimuli on a 17-inch colour computer monitor placed 60 cm from the eyes. The experiment was conducted on a PC computer using in-house software.
The two main variables were configuration (Old versus New) and set (1–4). The search task consisted of 24 blocks of 16 trials. Blocks were grouped into four sets, increasing the statistical power. Each block contained eight Old trials and eight New trials. The Old block of stimuli consisted of eight randomly generated configurations which were repeated throughout the entire experiment, once per set. The direction of the target rotation was chosen randomly, but it always appeared in the same location, with the same colour, within a particular configuration. The New configurations were generated randomly. The probability of chance repetition was extremely low, given the large number of unique configurations (17.4 billion) that could be generated from a sampling of 11 out of 47 locations (excluding the target location). The target consistently appeared in eight locations, in the same colour, throughout the experiment, as in the Old condition. This ruled out the possibility that the results were due to sensitivity to position frequency of the target. Thus target colour and location were preserved in both the Old and New conditions. Distractor locations in each configuration were sampled randomly from all possible locations. Target identity was randomized across trials to eliminate stimulus–response learning. The eccentricity or spatial location of the targets was chosen randomly and assigned to the two configuration conditions.
Analysis
Response times (RTs) of >10 s or <200 ms were discarded. Fewer than 1% of the trials were omitted by this trimming procedure. ERPs from correct trials were averaged into four sets. Each subject's ERPs from the set data were analysed separately and later combined in a within-subject analysis. ERP amplitude was calculated from baseline to peak, where baseline was the mean voltage in the 100 ms period preceding stimulus onset. Latency was calculated from stimulus onset at time 0 to peak.
Visual inspection showed that the primary difference between Old and New conditions occurred at ~200 ms. Thus peak amplitudes of short-latency ERPs that showed differentiation by condition were submitted to a t test between Old and New conditions. We defined this potential as any negative deflection showing modulation occurring between 150 and 250 ms. All electrode sites were inspected including sites in parietal and frontal regions. All differentiated ERPs were included in the analysis, regardless of the direction of change, i.e. whether the Old condition ERP amplitude was larger or smaller than the New condition amplitude. Data from adjacent electrode sites were not included to avoid the problem of correlated activity. Because each electrode strip contained 10 electrodes on average, thereby covering a large cortical area, multiple sites on one strip were included in the analysis as long as they were not adjacent. To ensure unbiased analysis of the data, ERPs recorded in set 4 were checked against those in set 3 to ensure that changes were consistent over time. Inconsistent results were assumed to be artefactual and were eliminated.
P100 and N100 sites were defined as any sharp positive (e.g. P100) or negative (e.g. N100) deflection peaking at 80–120 ms in V1 or V2. To assess possible contextual cueing-related changes in long-latency ERPs following P100 and N210, the area under the curve (AUC) was calculated (Allison et al., 1999) within a latency range of 200–400 ms, which captured most of this activity; measurements were pooled for sets 3 and 4 and submitted to t tests comparing condition.
|
Results |
|---|
|
TopAbstractIntroductionMethodResultsDiscussionAcknowledgementsReferences |
|---|
Behavioural performance
The contextual cueing effect is defined as an RT benefit in the Old condition compared with the New condition. Overall accuracy was 85%. Accuracy did not differ by condition, set, or their interaction (all F values <1).
The mean RTs for all correct trials within a set were computed separately for each condition, and were submitted to a repeated-measures ANOVA (analysis of variance) with configuration (Old versus New) and set (1–4) as factors. Overall search RT decreased over time [F(3,33) = 9.615, P < 0.0001] due to skill learning. A second learning component, particular to learning of the repeated configurations in the Old condition (the contextual cueing effect), was also found [F(1,11) = 47.998, P < 0.0001] (Fig. 2). The overall interaction was not significant, but a comparison across all blocks showed a significant interaction [F(23,253) = 2.139, P < 0.002], suggesting that performance was similar at the beginning of the experiment, but that learning of contexts rapidly ensued. There was no difference between conditions at set 1 [t(11) = 1.17, n.s.] but differences were found at sets 2–4 (all Ps <0 .01). By set 4, the cueing benefit was 149 ms.
|
ERPs
The ERP that consistently differentiated the New and Old conditions in sets 3 and 4 was a negative potential peaking at ~208 ms (termed N210) and defined as a sharp negativity with a peak latency between 150 and 250 ms. A representative recording is shown in Fig. 3A
. This ERP did not show any statistically significant differentiation by condition early in the experiment, at set 1, suggesting that the enhanced N210 in set 4 was indeed driven by contextual cueing and attentional training.
|
Over the patient group, there were 48 N210 sites; for statistical purposes, we excluded 13 sites on adjacent electrodes, which may have had correlated activity, leaving 35 sites in 12 patients. Contextual cueing (Old versus New) evoked a significantly larger N210 (P < 0.0001). N210s averaged across all patients (one site per patient with the largest condition-dependent amplitude difference, regardless of direction of change) are shown in Fig. 3B
. Contextual cueing did not affect peak latency (average latency =208.5 ± 34 ms in the New condition, 208.9 ± 30 ms in the Old condition) nor were there significant peak latency differences between anatomical regions.
The most prevalent N210 sites were in V2 (43%), although sites in V1 and in extrastriate cortex were also found (Table 1). All 48 sites that exhibited N210s are shown in views of the inferior and mesial surfaces of the brain (Fig. 4A–C). Although most of the patients had numerous electrodes in frontal, lateral temporal and parietal regions, there were no N210s or any other contextual cueing-related changes in ERPs at these sites.
|
|
To assess whether earlier ERPs were modified by contextual cueing, we searched for sites that generated N100 or P100 ERPs, which reflect initial activation of human V1 and V2 and are sensitive to low-level stimulus attributes such as luminance (Allison et al., 1999
). There were four P100 sites (Fig. 3C) and four N100 sites. There were no significant task-related effects on amplitude or latency for P100 or N100. However, a slow negative ERP following P100 was larger in the Old compared with the New condition (Fig. 3C); AUC measurements showed that Old amplitudes were significantly larger (P < 0.03) than New amplitudes in the 200–400 ms latency range (Fig. 3C, striped area). No significant AUC difference for the late negative ERP was found at N100 sites.
In six patients, a total of nine depth probes were placed to sample EEG activity in the hippocampus and adjacent temporal lobe structures. Post-operative MRIs showed that in all cases the end of the probe penetrated the hippocampus. None of these sites recorded N210s or other ERPs (Fig. 3D).
Two considerations suggest that the patients' epilepsy did not affect the ERPs recorded in this study: (i) in a larger sample of similar recordings, there were no N200 latency or amplitude differences between the normal and abnormal hemispheres (Allison et al., 1999); and (ii) all learning-related N210s and late negative ERPs were recorded from posterior temporal and occipital sites, whereas the epileptogenic foci were in anterior temporal and frontal regions. In three of the six patients with hippocampal depth probes, the anterior temporal lobe was determined to be the site of the epileptogenic focus; it is possible that ERPs in this region were affected in these patients.
|
Discussion |
|---|
|
TopAbstractIntroductionMethodResultsDiscussionAcknowledgementsReferences |
|---|
Contextual cueing of attention
It is commonly assumed that ERPs of increasing latency reflect activity in successive stages of a neural system. However, the major finding of this study is that contextual cueing modifies later ERPs generated in anatomically early areas including V1, the initial site of visual cortical processing. The modification consisted of an enhancement of N210 (Fig. 3A and B
), and an enhancement of a slow negative ERP at P100 sites, from 200 to 400 ms (Fig. 3C, striped area), 100–300 ms after the onset of initial activation of visual cortex (Fig. 3C). The `attentional amplification' (Posner and Dehaene, 1994) of N210 also occurred in several regions of extrastriate cortex (Fig. 4A–C). If these changes were due to modifications of early, bottom-up perceptual processes, modification of P100 and N100 (reflecting a portion of the initial activation of V1 and V2; Allison et al., 1999) would be expected. However, these ERPs were unaffected. In addition, a bottom-up effect independent of context would have to rely on low-level differences in Old and New displays, which seems highly unlikely since they were identical in all low-level features. The relatively late timing of these changes suggests that the mechanism of attentional guidance operates by an interaction between early and late visual areas. Re-entrant or feedback processes from higher areas may modulate processes in initial visual areas. Other groups have drawn similar conclusions based on the timing and location of word processing (Nobre et al., 1998), context encoding (e.g. Zipser et al., 1996; Gilbert, 1998), the allocation of spatial attention (Luck and Hillyard, 1994; Aine et al., 1995; Gratton, 1997; Vidyasasgar, 1998; Martinez et al., 1999) and intermodal selective attention (Mehta et al., 2000a, b; Foxe and Simpson, 2001). In an intracranial ERP study of attention to colours and words (Nobre et al., 1998), attention strongly modulated long-latency, but not short-latency, ERPs. The authors concluded that the modulation was due to top-down influences from downstream regions involved in word processing. However, in contrast to the present results, attention had the effect of reducing the amplitude of a long-latency negative ERP. The reason for this difference in attentional modulation is unknown but may be due to mechanisms specific to word processing.
In a recent study of spatial attention, modulation of V1 was detected by functional MRI, but evidence for delayed modulation was not present in scalp-recorded ERPs (Martinez et al., 1999). However, a follow-up study by the same group found some evidence for a late attentional affect in or around V1 (Martinez et al., 2001). The authors noted that `such a delayed attention effect . . . could have escaped detection if the striate cortex source were weak enough to be masked by the stronger sources that were concurrently active in extrastriate cortex.' Intracranial ERPs avoid this limitation by allowing direct recording from striate cortex, and demonstrate that a `delayed attention effect' is present in V1.
The results reported herein are also consistent with cognitive models which propose that context may influence visual processing within the first few hundred milliseconds of processing (Biederman et al., 1982). Although the ERP evidence clearly shows that learned context did not affect the initial volley of visual information into the cortex, the 200 ms latency of modulatory effects and their early anatomical foci support the hypothesis that top-down context information is affecting early perceptual processes. Understanding how such early contextual modulation influences the recognition of objects in scenes is an important issue for future research (Henderson and Hollingworth, 1999).
Sources of top-down modulation
We hypothesize that visual areas receive memory-based feedback from medial temporal lobe structures important for storing spatial context information (Chun and Phelps, 1999). Half of the patients in the present study had depth probes targeted to the hippocampus. These electrodes recorded either no ERPs (Fig. 3D) or small ERPs (0–20 µV) that were not modulated by contextual cueing. Except for large ERPs to infrequent target-related stimuli (McCarthy et al., 1989), most recordings from the human hippocampus have not demonstrated learning-related ERP changes (Nobre and McCarthy, 1995; Puce et al., 1999). Hence, the present results were not helpful in evaluating possible feedback effects from medial temporal lobe structures, and we do not interpret the null effect here as evidence against possible hippocampal involvement in contextual cueing tasks.
Mehta and colleagues (Mehta et al., 2000a, b) recorded from visual cortex while monkeys performed an attention task. They found early attentional modulation in area V4 and late modulation in area V1, suggesting that V4 might be a source of feedback to V1. While this is a plausible source of feedback insofar as V1 is concerned, we also found modulation of N210 and later activity in regions of anterior extrastriate cortex (Fig. 4) and in the posterior fusiform gyrus in which the possible human homologue of monkey area V4 is primarily located (Clarke and Miklossy, 1990; Allison et al., 1993; Sereno et al., 1995; De Yoe et al., 1996). The source of feedback to these regions of extrastriate cortex is unknown, but might arise from the human homologues of areas TEO and TE, which send feedback projections to earlier visual areas including V4 (Rockland and Drash, 1996).
Contextual cueing differs from repetition priming
Although our task involved repetition of configurations, contextual cueing can be distinguished behaviourally from repetition priming (Chun and Jiang, 1998). First, Chun and Jiang showed that the specific form of the items making up the Old configurations can be changed midway through the experiment with no detrimental affect on contextual learning. Thus contextual cueing is independent of the surface features of the array, which is atypical of low-level repetition priming. Secondly, Chun and Jiang showed that repeated configurations did not benefit search performance when the contexts and target locations were decorrelated. Rather, subjects learned associations between configurations and embedded target locations.
Several intracranial ERP studies also suggest that the N210 effect found in this study is not due to repetition priming (reviewed by Guillem et al., 1999) or semantic priming (reviewed by Schacter and Buckner, 1998). In a habituation study, the first repetition of the stimulus significantly decreased N200 amplitude to face stimuli (Puce et al., 1999), while semantic priming had no significant effect on word-related (Nobre et al., 1994) or face-related (Puce et al., 1999) N200 amplitude. Repetition and semantic priming also significantly reduced the amplitude of N400 and N700 ERPs (Nobre and McCarthy, 1995; Guillem et al., 1999; Puce et al., 1999) recorded from the temporal lobe. Thus repetition and semantic priming decrease or have no effect on the amplitude of negative ERPs, results opposite to the enhancement of negative ERPs seen in this study.
Relationship to N2pc
Several studies have identified a scalp-recorded ERP component in visual search tasks that may be related to attentional filtering. This ERP is called the `N2pc' (e.g. Luck and Hillyard, 1994b). It consists of a negative deflection that occurs 200–300 ms following target onset. N2pc is localized to the occipital region and is evoked by targets or difficult non-targets embedded in an array of distractor items that need to be filtered in order to discriminate the target correctly. N2pc is a slow negative-going shift without a discrete peak; it is not clear whether N210, the late negative ERP or both are the cortical surface-recorded counterparts of N2pc. It is possible that all three ERPs reflect a common underlying mechanism, providing additional evidence for our conclusion above that contextual modification of N210 and the late negative ERP reflect an attentional rather than a priming effect.
Physiological mechanisms of top-down modulation
The neural mechanisms responsible for the postulated feedback processes are unknown, but physiological and anatomical data allow an informed guess. Surface-negative ERPs are thought to reflect excitation of the apical dendrites of pyramidal cells (Schlag, 1973; Wood and Allison, 1981). Anatomical studies demonstrate that feedback projections are primarily to superficial (layer I) or deep (layers V–VI) cortex (Rockland, 1997). The superficial projection, which is mainly to the apical dendrites of pyramidal cells, would be expected to influence dendritic ERPs such as N210. In rat V1, input to layer 1 (Cauller and Conners, 1994), including feedback projections to layer 1 (Shao and Burkhalter, 1996; Nowak et al., 1997), evokes long-duration depolarizing potentials. ERPs, multi-unit activity and current source density in several areas of visual cortex have been analysed while monkeys performed a visual attention task (Mehta et al., 2000a, b). Attentional modulation of neuronal activity in V1 was found to be due to depolarizing current sinks in superficial layers (I and II), in agreement with our inference in humans. The modulation in V1 occurred at 200–400 ms, similar to the modulation we observed in V1 at P100 sites (Fig. 3C, striped area). Taken together, the human, rat and monkey recordings provide strong evidence for late attentional and contextual excitatory modulation of anatomically early areas of visual cortex, and provide a plausible neurophysiological explanation of its mechanism of action.
|
Acknowledgements |
|---|
|
TopAbstractIntroductionMethodResultsDiscussionAcknowledgementsReferences |
|---|
We wish to thank J. Jasiorkowski for assistance, and Dr D. Spencer and the staff of the Yale Epilepsy Surgery Program for their cooperation in the recordings described here. We also wish to thank S. Hillyard for helpful comments on earlier versions of this paper. This work was supported by the Veterans Administration, by NIMH grant MH-05286, NSF grant BCS-9817349, and by an NSF predoctoral fellowship to I.R.O.
|
References |
|---|
|
TopAbstractIntroductionMethodResultsDiscussionAcknowledgementsReferences |
|---|
Aine CJ, Supek S, George JS. Temporal dynamics of visual-evoked neuromagnetic sources: effects of stimulus parameters and selective attention. Int J Neurosci 1995; 80: 79–104.[Medline]
Allison T, Begleiter A, McCarthy G, Roessler E, Nobre AC, Spencer DD. Electrophysiological studies of color processing in human visual cortex. Electroencephalogr Clin Neurophysiol 1993; 88: 343–55.[Web of Science][Medline]
Allison T, Puce A, Spencer DD, McCarthy G. Electrophysiological studies of human face perception. I: potentials generated in occipitotemporal cortex by face and non-face stimuli. Cereb Cortex 1999; 9: 415–30.
Biederman I, Mezzanotte RJ, Rabinowitz JC. Scene perception: detecting and judging objects undergoing relational violations. Cognit Psychol 1982; 14: 143–77.[Web of Science][Medline]
Bravo M J, Nakayama K. The role of attention in different visual-search tasks. Percept Psychophys 1992; 51: 465–72 .[Web of Science][Medline]
Cauller LJ, Connors BW. Synaptic physiology of horizontal afferents to layer 1 in slices of rat SI neocortex. J Neurosci 1994;14: 751–62.[Abstract]
Chun MM. Contextual cueing of visual attention. Trends Cogn Sci 2000; 4: 170–8.[Web of Science][Medline]
Chun MM, Jiang Y. Contextual cueing: implicit learning and memory of visual context guides spatial attention. Cognit Psychol 1998; 36: 28–71.[Web of Science][Medline]
Chun MM, Phelps EA. Memory deficits for implicit contextual information in amnesic subjects with hippocampal damage. Nat Neurosci 1999; 2: 844–7.[Web of Science][Medline]
Clarke S, Miklossy J. Occipital cortex in man: organization of callosal connections, related myelo- and cytoarchitecture, and putative boundaries of functional visual areas. J Comp Neurol 1990; 298: 188–214.[Web of Science][Medline]
De Yoe EA, Carman GJ, Bandettini P, Glickman S, Wieser J, Cox R, et al. Mapping striate and extrastriate visual areas in human cerebral cortex. Proc Natl Acad Sci USA 1996; 93: 2382–6.
Foxe JJ, Simpson GV. Timecourse of activation flow from V1 to frontal cortex in humans: a framework for defining `early' visual processing. Neuroimage. In press 2001.
Gandhi S P, Heeger DJ, Boynton GM. Spatial attention affects brain activity in human primary visual cortex. Proc Natl Acad Sci USA 1999; 96: 3314–19.
Gilbert CD. Adult cortical dynamics. [Review]. Physiol Rev 1998; 78: 467–85.
Gratton G. Attention and probability effects in the human occipital cortex: an optical imaging study. Neuroreport 1997; 8: 1749–53.[Web of Science][Medline]
Guillem F, Rougier A, Claverie B. Short- and long-delay intracranial ERP repetition effects dissociate memory systems in the human brain. J Cogn Neurosci 1999; 11: 437–58.[Web of Science][Medline]
Halgren E, Squires NK, Wilson CL, Rohrbaugh JW, Babb TL, Crandall PH. Endogenous potentials generated in the human hippocampal formation and amygdala by infrequent events. Science 1980; 210: 803–5.
Heinze HJ, Mangun GR, Burchert W, Hinrichs H, Scholz M, Munte TF, et al. Combined spatial and temporal imaging of brain activity during visual selective attention in humans. Nature 1994; 372: 543–6.[Medline]
Henderson JM, Hollingworth A. High-level scene perception. [Review]. Annu Rev Psychol 1999; 50: 243–71.[Web of Science][Medline]
Hillyard SA, Anllo-Vento L. Event-related brain potentials in the study of visual selective attention. [Review]. Proc Natl Acad Sci USA 1998; 95: 781–7.
Ito M, Gilbert CD. Attention modulates contextual influences in the primary visual cortex of alert monkeys. Neuron 1999; 22: 593–604.[Web of Science][Medline]
Johnston WA, Hawley KJ, Plewe SH, Elliott JM, DeWitt MJ. Attention capture by novel stimuli. J Exp Psychol Gen 1990; 119: 397–411.[Web of Science][Medline]
Kanwisher N, McDermott J, Chun MM. The fusiform face area: a module in human extrastriate cortex specialized for face perception. J Neurosci 1997; 17: 4302–11.
Kastner S, De Weerd P, Desimone R, Ungerleider LG. Mechanisms of directed attention in the human extrastriate cortex as revealed by functional MRI. Science 1998; 282: 108–11.
Luck SJ, Hillyard SA. Electrophysiological correlates of feature analysis during visual search. Psychophysiology 1994a; 31: 291–308.[Web of Science][Medline]
Luck SJ, Hillyard SA. Spatial filtering during visual search: evidence from human electrophysiology. J Exp Psychol Hum Percept Perform 1994b; 20: 1000–14.[Web of Science][Medline]
Luck SJ, Chelazzi L, Hillyard SA, Desimone R. Neural mechanisms of spatial selective attention in areas V1, V2, and V4 of macaque visual cortex. J Neurophysiol 1997; 77: 24–42.
Mangun GR, Hopfinger JB, Kussmaul CL, Fletcher EM, Heinze H-J. Covariations in ERP and PET measures of spatial selective attention in human extrastriate visual cortex. Hum Brain Mapp 1997; 5: 273–9.[Web of Science]
Mangun GR, Buonocore MH, Girelli M, Jha AP. ERP and fMRI measures of visual spatial selective attention. Hum Brain Mapp 1998; 6: 383–9.[Web of Science][Medline]
Martinez A, Anllo-Vento L, Sereno MI, Frank LR, Buxton RB, Dubowitz DJ, et al. Involvement of the striate and extrastriate visual cortical areas in spatial attention. Nat Neurosci, 1999; 2: 364–9.[Web of Science][Medline]
Martinez A, DiRusso F, Anllo-Vento L, Sereno MI, Buxton RB, Hillyard, SA. Putting spatial attention on the map: timing and localization of stimulus selection processes in striate and extrastriate visual areas. In press 2001.
McCarthy G, Wood CC, Williamson PD, Spencer DD. Task-dependent field potentials in human hippocampal formation. J Neurosci 1989; 9: 4253–68.[Abstract]
McCarthy G, Spencer DD, Riker RJ. The stereotaxic placement of depth electrodes. In: Luders H, editor. Epilepsy surgery. New York: Raven Press; 1991. p. 385–93.
Mehta AD, Ulbert I , Schroeder CE. Intermodal selective attention in monkeys. I: distribution and timing of effects across visual areas. Cereb Cortex 2000a; 10: 343–58.
Mehta AD, Ulbert I, Schroeder CE. Intermodal selective attention in monkeys. II: physiological mechanisms of modulation. Cereb Cortex 2000b; 10: 359–70.
Miller J. Components of the location probability effect in visual search tasks. J Exp Psychol Hum Percept Perform 1998; 14: 453–71.
Nobre AC, McCarthy G. Language-related field potentials in the anterior-medial temporal lobe. II: effects of word type and semantic priming. J Neurosci 1995; 15: 1090–8.[Abstract]
Nobre AC, Allison T, McCarthy G. Word recognition in the human inferior temporal lobe. Nature 1994; 372: 260–3.[Medline]
Nobre AC, Allison T, McCarthy G. Modulation of human extrastriate visual processing by selective attention to colours and words. Brain 1998; 121: 1357–68.
Nowak LG, James AC, Bullier J. Corticocortical connections between visual areas 17 and 18a of the rat studied in vitro: spatial and temporal organisation of functional synaptic responses. Exp Brain Res 1997; 117: 219–41.[Web of Science][Medline]
Posner MI, Dehaene S. Attentional networks. [Review]. Trends Neurosci 1994; 17: 75–9.[Web of Science][Medline]
Puce A, Allison T, McCarthy G. Electrophysiological studies of human face perception. III: effects of top-down processing on face-specific potentials. Cereb Cortex 1999; 9: 445–58.
Rockland KS. Elements of cortical architecture: hierarchy revisited. In: Rockland KS, Kaas JH, Peters A, editors. Cerebral cortex, Vol. 12: extrastriate cortex in primates. New York: Plenum Press; 1997. p. 243–93.
Rockland KS, Drash GW. Collateralized divergent feedback connections that target multiple cortical areas. J Comp Neurol 1996; 373: 529–48.[Web of Science][Medline]
Roelfsema PR, Lamme VA, Spekreijse H. Object-based attention in the primary visual cortex of the macaque monkey. Nature 1998; 395: 376–81.[Medline]
Schacter DL, Buckner RL. Priming and the brain. [Review]. Neuron 1998; 20: 185–95.[Web of Science][Medline]
Schlag J. Generation of brain evoked potentials. In: Thompson RF, Patterson MM, editors. Bioelectric recording techniques, Part A: cellular processes and brain potentials. New York: Academic Press; 1973. p. 273–316.
Schmolesky MT, Wang Y, Hanes DP, Thompson KG, Leutgeb S, Schall JD, et al. Signal timing across the macaque visual system. J Neurophysiol 1998; 79: 3272–8.
Sereno MI, Dale AM, Reppas JB, Kwong KK, Belliveau JW, Brady TJ, et al. Borders of multiple visual areas in humans revealed by functional magnetic resonance imaging. Science 1995; 268: 889–93.
Shao Z, Burkhalter A. Different balance of excitation and inhibition in forward and feedback circuits of rat visual cortex. J Neurosci 1996; 16: 7353–65.
Shaw ML. A capacity allocation model for reaction time. J Exp Psychol Hum Percept Perform 1978; 4: 586–98.[Web of Science]
Shaw ML , Shaw P. Optimal allocation of cognitive resources to spatial locations. J Exp Psychol Hum Percept Perform 1977; 3: 201–11.[Web of Science][Medline]
Somers DC, Dale AM, Seiffert AE, Tootell RB. Functional MRI reveals spatially specific attentional modulation in human primary visual cortex. Proc Natl Acad Sci USA 1999; 96: 1663–8.
Spencer SS, Spencer DD, Williamson PD, Mattson RH. The localizing value of depth electroencephalography in 32 patients with refractory epilepsy. Ann Neurol 1982; 12: 248–53.[Web of Science][Medline]
Talairach J, Tournoux P. Co-planar stereotaxic atlas of the human brain. Stuttgart: Thieme; 1988.
Treisman A , Gormican S. Feature analysis in early vision: evidence from search asymmetries. Psychol Rev 1988; 95: 15–48.[Web of Science][Medline]
Vidyasagar TR. Gating of neuronal responses in macaque primary visual cortex by an attentional spotlight. Neuroreport 1998; 9: 1947–52.[Web of Science][Medline]
Wang Q, Cavanagh P, Green M. Familiarity and pop-out in visual search. Percept Psychophys 1994; 56: 495–500.[Web of Science][Medline]
Watanabe T, Harner AM, Miyauchi S, Sasaki Y, Nielsen M, Palomo D, et al. Task-dependent influences of attention on the activation of human primary visual cortex. Proc Natl Acad Sci USA 1998; 95: 11489–92.
Wood CC, Allison T. Interpretation of evoked potentials: a neurophysiological perspective. Can J Psychol 1981; 35: 113–35.[Web of Science][Medline]
Yantis S, Jonides J. Abrupt visual onsets and selective attention: evidence from visual search. J Exp Psychol Hum Percept Perform 1984; 10: 601–21.[Web of Science][Medline]
Zipser K, Lamme VA, Schiller PH. Contextual modulation in primary visual cortex. J Neurosci 1996; 16: 7376–89.
Received August 30, 2000. Revised January 20, 2001. Accepted March 12, 2001.
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  F. Di Russo, T. Aprile, G. Spitoni, and D. Spinelli Impaired visual processing of contralesional stimuli in neglect patients: a visual-evoked potential study Brain, March 1, 2008; 131(3): 842 - 854. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. Murayama, E. Perez-Garci, H.-R. Luscher, and M. E. Larkum Fiberoptic System for Recording Dendritic Calcium Signals in Layer 5 Neocortical Pyramidal Cells in Freely Moving Rats J Neurophysiol, September 1, 2007; 98(3): 1791 - 1805. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. E. Larkum, W. Senn, and H.-R. Luscher Top-down Dendritic Input Increases the Gain of Layer 5 Pyramidal Neurons Cereb Cortex, October 1, 2004; 14(10): 1059 - 1070. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
F. Di Russo, A. Martinez, and S. A. Hillyard Source Analysis of Event-related Cortical Activity during Visuo-spatial Attention Cereb Cortex, May 1, 2003; 13(5): 486 - 499. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. W. Spratling Cortical region interactions and the functional role of apical dendrites. Behav Cogn Neurosci Rev, September 1, 2002; 1(3): 219 - 228. [Abstract] [PDF]
|
|
|
|
|
|
F. Castelli, C. Frith, F. Happe, and U. Frith Autism, Asperger syndrome and brain mechanisms for the attribution of mental states to animated shapes Brain, August 1, 2002; 125(8): 1839 - 1849. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||