Hemispheric asymmetries for visual and auditory temporal processing: an evoked potential study

Abstract

Lateralization for temporal processing was investigated using evoked potentials to an auditory and visual gap detection task in 12 dextral adults. The auditory stimuli consisted of 300-ms bursts of white noise, half of which contained an interruption lasting 4 or 6 ms. The visual stimuli consisted of 130-ms flashes of light, half of which contained a gap lasting 6 or 8 ms. The stimuli were presented bilaterally to both ears or both visual fields. Participants made a forced two-choice discrimination using a bimanual response. Manipulations of the task had no effect on the early evoked components. However, an effect was observed for a late positive component, which occurred approximately 300–400 ms following gap presentation. This component tended to be later and lower in amplitude for the more difficult stimulus conditions. An index of the capacity to discriminate gap from no-gap stimuli was gained by calculating the difference waveform between these conditions. The peak of the difference waveform was delayed for the short-gap stimuli relative to the long-gap stimuli, reflecting decreased levels of difficulty associated with the latter stimuli. Topographic maps of the difference waveforms revealed a prominence over the left hemisphere. The visual stimuli had an occipital parietal focus whereas the auditory stimuli were parietally centered. These results confirm the importance of the left hemisphere for temporal processing and demonstrate that it is not the result of a hemispatial attentional bias or a peripheral sensory asymmetry.

Introduction

The first insights into functional specialization of the cerebral hemispheres came from studies investigating the effects of unilateral lesions. Dax, Broca and Wernicke observed that lesions to the left hemisphere (LH) gave rise to deficits associated with the production and comprehension of speech. Subsequently, Hughlings-Jackson reasoned that, if the LH was specialized for language, then the right hemisphere (RH) must be specialized for perception (Corballis, 1991). These early descriptions of cerebral function adopted a top-down approach to cerebral asymmetry, which emphasized psychological functions and where they were localized in the brain. With the advent of cognitive psychology and computer modeling, there has been a trend towards a bottom-up approach that describes hemispheric function in terms of the processes that underlie functional specialization. In other words, there has been a move away from describing the functions of the cerebral hemispheres towards an approach that identifies the processes that are lateralized.

There are a number of reasons why an information processing approach to describing cerebral asymmetry may be particularly useful. First, descriptions of cerebral asymmetry based on an information processing approach are able to account for many of the discrepancies in laterality research that top-down approaches cannot. For example, simply describing the LH as ‘verbal’ and the RH as ‘non-verbal’ cannot account for the left and right hemisphere advantages observed for the recognition of Kana and Kanji, respectively (Abe and Yokoyama, 1994). Similarly, a dissociation between the hemispheres for the recognition of the local and global features of a face (Patterson and Bradshaw, 1975) does not fit with top-down approaches to cerebral asymmetry. These results, however, can be explained by information processing descriptions. In both cases, LH advantages arise when the information is broken down into parts and processed in an analytic, serial fashion. In contrast, RH advantages emerge when the information is processed holistically (see Bradshaw and Nettleton, 1981 for more details).

Information processing approaches to cerebral asymmetry also tie in with current theories of how the hominid brain may have evolved. There is increasing evidence that cerebral asymmetry has not suddenly emerged in humans and that it can be seen in a number of other species. For example, right hand preferences have been shown in chimpanzees (Hopkins and Fernández-Carriba, 2000), LH specialization for species-specific calls in macaques (Heffner and Heffner, 1984) and RH specialization for emotional expression (Hauser, 1993) and face recognition (Vermeire et al., 1998) in old-world monkeys. On the basis of results such as these, a number of researchers have suggested that human cerebral asymmetry reflects a continuous thread of hemispheric specialization that stretches back into our past and which can be seen in other species (Bradshaw and Rogers, 1993, Corballis, 1989). What characterizes this cerebral specialization are modes of cognition rather than language per se. Corballis (1989) emphasized the praxic, time-dependent properties of the LH whereas Bradshaw and Rogers (1993) stressed RH specialization for spatial cognition.

One recent description of cerebral asymmetry, which is based on the information processing approach, has suggested that the hemispheres are differentially specialized for temporal processing (Miller, 1996, Nicholls, 1996). Temporal processing is taken to describe a process where stimuli are distinguished from one another through time-dependent properties that range in duration from a few milliseconds through to 200 ms. This type of processing is thought to occur at a relatively low-level of cognition and may underlie the operation of higher level processes such as language (Belin et al., 1998, Miller, 1996, Nicholls, 1996).

Unilateral lesion research generally supports a LH advantage for temporal processing. Von Steinbüchel et al. (1999) required patients with unilateral anterior or posterior lesions to determine the order (right or left) for pairs of binaurally presented clicks. Von Steinbüchel et al. (1999) found that only patients with posterior lesions to the LH were impaired in their ability to determine temporal order. Lackner and Teuber (1973) have reported similar performance decrements in patients with LH lesions for two-click fusion thresholds. The middle portion of the superior temporal gyrus of the left hemisphere may play a particularly important role in auditory temporal processing. Damage to this region can result in pure word deafness, a condition that impairs the recognition of the sounds of speech. Phillips and Farmer (1990) suggested that pure word deafness stems from an inability to process rapidly changing acoustic stimuli. In support of this proposition, they found that patients with word deafness were unable to recognize fast-changing stop consonants (e.g. ‘b’) whereas their capacity to discriminate slow-changing steady-state vowels (e.g. ‘e’) remained intact. When combined, these results suggest that, not only is the LH involved in auditory temporal processing, but also that the temporal/posterior regions appear more important in this type of processing.

The LH advantage for auditory temporal processing appears to extend to the visual modality. Goldman et al. (1968) compared critical flicker fusion thresholds between patients with left or right temporal lobectomies and a group of normal controls. Patients with LH temporal lobectomies were significantly impaired compared to the matched control group. A similar LH advantage for sequencing visual and auditory events has been reported by Efron (1963a). He required groups of aphasic patients with LH damage and patients with RH damage to make temporal order judgements of two visual or auditory events. Patients with LH damage performed worse than did the RH group. These results have been replicated in a study conducted by Swisher and Hirsh (1972) which also examined temporal ordering deficits within the visual and auditory modalities. They found that patients with LH temporal lesions performed worse compared to the RH lesion group (though, see Aram and Ekelman, 1988 for opposite results). These results demonstrate a LH temporal processing advantage for visual stimuli. Not only do these functions appear to be located in the LH, but also, they appear to be located in the temporal lobe, along with auditory temporal processing functions.

Temporal processing asymmetries have also been observed in non-clinical populations. For the visual modality, experimenters have used divided visual field techniques to deliver stimuli unilaterally to the hemispheres. Research using bilateral presentation procedures has generally yielded a ‘right first’ advantage for simultaneity or temporal order judgements (Corballis, 1996, Efron, 1963b, St. John, 1998). This ‘right first’ advantage is usually thought to reflect the direct access the right visual field has to the temporal processing mechanisms located in the LH. For trials which start on the left, the first stimulus within the pair is delayed by the time it takes to cross the corpus callosum from the RH to the LH. This delay effectively reduces the subjective delay between the stimuli, rendering temporal discrimination more difficult (Mills and Rollman, 1980). A right visual field advantage has also been observed for unilaterally presented stimuli. Examples of tasks that have yielded a right visual field advantage include, inspection time (Elias et al., 1999, Nicholls and Atkinson, 1993, Nicholls and Cooper, 1991), judgements of duration (Grondin, 1998), simultaneity judgements (Nicholls, 1994) and two-flash fusion (Nicholls, 1994). It should be noted, however, that support for a right visual field advantage for temporal processing advantage is not unanimous. Sadler and Deary (1996) found no asymmetry for an inspection time task, Heider and Groner (1997) found a material-specific effect for words and faces, but no asymmetry for the effect of masking, and Forster et al. (2000) found a left visual field advantage for simultaneity judgements in split brain subjects.

For the auditory modality, hemispheric asymmetries have been investigated by comparing performance between the ears. Using bilateral presentations, Mills and Rollman (1980) reported an advantage when the first member within a pair of clicks was delivered to the right ear for a temporal order judgement task. A right ear advantage has also been reported for a variety of tasks that use unilaterally presented stimuli. Examples of these tasks include gap detection in bursts of noise (Brown and Nicholls, 1997, Kelso et al., 2000, Nicholls et al., 1999, Vroon et al., 1977; though see Efron et al., 1985 for negative results), recognition of Morse-code sequences (Papcun et al., 1974) and the detection of temporal order (Cutting, 1974, Divenyi and Efron, 1979).

A number of neuro-structural models have been developed which can account for the LH temporal processing advantage. Miller (1996) proposed that the LH contained more slow conducting, unmyelinated axons with conduction times in excess of 100 ms. Slow neural conduction within the LH was proposed to prevent the neural integration of events defined by periods of time lasting between 10 and 100 ms, thus allowing the LH to represent rapid temporal events. In contrast, the RH was proposed to contain more fast-conducting, myelinated axons with conduction times less that 10 ms. As a result, stimuli defined by periods lasting between 10 and 100 ms will be integrated and lost in the RH. Another model proposed by Nicholls (1996) has attributed temporal processing asymmetries to the intra-regional and inter-regional neural organization of the left and right hemispheres, respectively (see Goldberg and Costa, 1981 and Gur et al., 1980 for evidence of such an asymmetry). Nicholls (1996) proposed that the inter-regional organization of the RH increased the level of neural feedback and feedforward and that this ‘blurred’ the temporal signal. In contrast, the intra-regional organization of the LH was proposed to reduce the blurring of the signal, resulting in a more accurate representation of the temporal properties of a stimulus. One advantage of Nicholls’ (1996) model is that it predicts a LH temporal processing advantage for stimuli defined by periods less than and greater than 10 ms. Thus, it can account for studies that have shown a LH advantage for visual simultaneity judgments with onset asynchronies lasting between 10 and 25 ms (Nicholls, 1994) and for auditory gap detection tasks with gaps lasting between 2 and 8 ms (Brown and Nicholls, 1997).

While the evidence in support of a temporal processing asymmetry may appear relatively strong, a number of researchers have argued against such an asymmetry. Geffen et al. (1996) suggested that the right visual field and ear advantages reflected attentional asymmetries between the hemispaces rather than a functional hemispheric asymmetry. Research supports the proposition that perceptual asymmetries can be affected by shifts in auditory (Mondor and Bryden, 1992) and visual (Nicholls and Wood, 1998) attention. In relation to temporal processing asymmetries, Geffen et al. (1996) found a right hand advantage for tactile simultaneity judgements when the left and right hands were placed in their respective hemispaces, but not when the hands were placed along the midline. On the basis of these results, Geffen et al. (1996) suggested that the right-hand advantage observed in the lateral hand placement condition was the result of a general rightward attentional bias rather than a LH specialization for temporal processing. However, Nicholls and Whelan (1998) have challenged this conclusion. They required participants to detect short interruptions in vibrations delivered unilaterally to the hands. The effect of hemispace was manipulated using lateral and midline hand placements. A right hand advantage was found for the task, and although the asymmetry was accentuated in the lateral hand-placement condition, it was still present when the hands were placed along the midline. Nicholls and Lindell (2000) have reported similar results. They required participants to make simultaneity judgments for pairs of vibrations delivered unilaterally to the hands. The effect of hemispace was manipulated using ipsilateral and contralateral (arms crossed) hand placement conditions. Irrespective of hand placement, a right hand advantage was observed for simultaneity judgements.

While the studies conducted by Nicholls and Lindell (2000) and Nicholls and Whelan (1998) demonstrate that hemispatial biases do not account for the temporal processing asymmetry, they do not rule out the possibility that the right hand is simply more sensitive. A similar argument could be made for stimuli delivered to the right ear or to the right visual field. In both cases, the perceptual asymmetry reflects asymmetries at a peripheral/sensory level rather than at a central/cortical level. In support of this proposition, Khalfa et al. (1998) investigated transient evoked otoacoustic emissions for the medial olivocochlear nucleus. They found evidence for increased auditory information reception in the right ear compared to the left. Thus, the right ear advantage found for temporal processing (e.g. Brown and Nicholls, 1997) may reflect a peripheral asymmetry at the level of the medulla rather than a cerebral asymmetry.

One way in which hemispheric asymmetries in temporal processing ability can be assessed more directly is through scanning techniques. To this end, Belin et al. (1998) used positron emission topography to assess the contribution of the hemispheres to temporal processing. Participants listened to non-verbal sounds containing either rapid (40-ms) or extended (200-ms) frequency transitions. For the slower transitions, cerebral activation was almost symmetrical. For rapid transitions, however, activation was significantly higher in the superior temporal gyri and dorsolateral prefrontal regions of the LH when compared to the corresponding regions of the RH. A LH focus for temporal processing has also been reported by Nicholls et al. (1999) using electroencephalographic (EEG) techniques. For an auditory gap detection task, they found higher levels of beta activity over the left temporal lobe than over the corresponding area in the RH. They concluded that this asymmetry in beta activity reflected a higher level of brain activity in the left temporal lobe for the gap detection task.

The research conducted by Belin et al. (1998) and Nicholls et al. (1999) both suggest a left hemisphere focus for temporal processing. The present study sought to build upon this research using an evoked potential study of temporal processing. There are a number of potential advantages of evoked potential research. First, because evoked potentials have a high degree of temporal precision, they provide an insight into the temporal dynamics of processing. Second, because relatively large numbers (in this case, 32) of electrodes can be attached, an insight into the spatial dynamics and localization of temporal processing can be gained.

Evoked potentials were measured while participants performed gap detection tasks in the auditory or visual modalities. The auditory gap detection task was based on one used by Brown and Nicholls (1997) and required participants to detect a brief interruption lasting either 4 or 6 ms located in the middle of a burst of white noise lasting 300 ms. The visual gap detection task was based on one used by Nicholls (1994) and required participants to detect an interruption lasting between 6 and 8 ms in a flash of light lasting 130 ms. Brown and Nicholls (1997) and Nicholls (1994) have demonstrated substantial right ear and right visual field advantages for auditory and visual gap detection tasks, respectively. The right visual field and right ear advantages were reflected in lower levels of error, faster reaction times (RT) and lower levels of response bias towards ‘no-gap’ responses.

Given that the clinical research has consistently implicated the temporal lobes in temporal processing, we expected asymmetries in the evoked potentials to be most apparent in this region. Visual and auditory tasks were included in this study to explore the proposition that the temporal processing ability of the temporal lobe is multimodal. If this proves to be the case, evoked potential maps should be similar for the visual and auditory tasks. It is not expected that asymmetries should arise for components that appear prior to the presentation of the gap stimulus. However, components that appear following gap presentation may be associated with the decision process, and therefore may demonstrate asymmetries. Particular attention was paid to late positive components, occurring at approximately 300–400 ms, which are thought to reflect processing associated with task difficulty (Palmer et al., 1994). Ability to process the stimuli was evaluated by generating difference waveforms between the gap and no-gap stimuli. If the LH is more able to discriminate between these stimuli, then it follows that the evoked potentials to these stimuli should be maximally divergent over the LH.