A magnetic stimulation examination of orthographic neighborhood effects in visual word recognition

The split-fovea theory proposes that visual word recognition is mediated by the splitting of the foveal image, with letters to the left of fixation projected to the right hemisphere (RH) and letters to the right of fixation projected to the left hemisphere (LH). We applied repetitive transcranial magnetic stimulation (rTMS) over the left and right occipital cortex during a lexical decision task to investigate the extent to which word recognition processes could be accounted for according to the split-fovea theory. Unilateral rTMS significantly impaired lexical decision latencies to centrally presented words, supporting the suggestion that foveal representation of words is split between the cerebral hemispheres rather than bilateral. Behaviorally, we showed that words that have many orthographic neighbors sharing the same initial letters ("lead neighbors") facilitated lexical decision more than words with few lead neighbors. This effect did not apply to end neighbors (orthographic neighbors sharing the same final letters). Crucially, rTMS over the RH impaired lead-, but not end-neighborhood facilitation. The results support the split-fovea theory, where the RH has primacy in representing lead neighbors of a written word.     

Hemispheric semantic priming in the single word presentation task

Functional brain asymmetries in semantic activation were studied by presenting categorically related (e.g. TABLE-BED) or unrelated primes and targets to the left visual field (LVF)/right hemisphere or to the right visual field (RVF)/left hemisphere in the single word presentation lexical decision task. The results showed that the primes in the RVF/left hemisphere primed lexical decisions to the targets both in the RVF and in the LVF. However, the primes in the LVF/right hemisphere did not induce any priming in the LVF or RVF. These results suggest that the left hemisphere automatically activates categorically related meanings in both hemispheres. The role of the right hemisphere in automatic semantic processing may be very limited.     

Hemispheric asymmetries in memory processes as measured in a false recognition paradigm

Although memory differs in important ways between the left and right cerebral hemispheres, the nature of these differences remains controversial. We examined this issue in two experiments using a false memory paradigm that allowed novel tests of two theories that have not been assessed in a common paradigm previously. Lists of semantically related words (e.g., bed, rest, wake…), all highly associated to one “critical” word (e.g., sleep), were presented auditorily during a study phase. Memory for both the related words and the critical words was measured in a subsequent old/new recognition test using divided-visual- field word presentations. The most important results were that the ability to correctly reject previously unpresented words was greater when test items were presented to the right visual field/left hemisphere (RVF/LH) than to the left visual field/right hemisphere (LVF/RH) and that participants were more confident in correctly rejecting unpresented words when test items were presented to the RVF/LH than to the LVF/RH. Results were in line with the theory that associative activation of semantic information is restricted in the left hemisphere but diffuse in the right; however, these results contrasted with the theory that memory traces are interpretive in the left hemisphere but veridical in the right. A potential resolution to the seemingly contradictory theories of asymmetries in memory processing is briefly discussed.     

Interhemispheric cooperation for familiar but not unfamiliar face processing

Evidence for interhemispheric cooperation during language processing has been demonstrated for words, but not for meaningless pseudowords. Specifically, responses were found to be faster and more accurate when identical copies of a word were presented bilaterally to both hemispheres, relative to unilateral single presentations. This bilateral advantage for words seems to be a robust effect in normals. The present study addressed the question of whether the bilateral advantage is restricted to lexical material or whether it is a more global phenomenon occurring for meaningful material in general. Thirty healthy participants performed a familiarity decision in which one copy of familiar and unfamiliar faces was presented tachistoscopically to the right visual hemifield (RVF), the left visual hemifield (LVF) or simultaneously to both visual hemifields (bilateral condition, BVF). We obtained a highly significant familiarity by visual field interaction(P < 0.006) showing that only for familiar faces, a bilateral advantage was obtained. Unfamiliar face processing did not yield a bilateral advantage. We conclude that interhemispheric cooperation only occurs for meaningful material for which cortical representations can be assumed.     

Orthographic and phonological priming in the two cerebral hemispheres

The patterns of activation invoked in the two cerebral hemispheres by written words may be different. Two lexical decision experiments investigated several aspects of such activation patterns. Experiment 1 tested phonological and orthographic priming in the hemispheres, manipulating two levels of phonological and two levels of orthographic similarity. Orthographic priming in the left visual field (LVF) was significantly larger than in the right visual field (RVF). In Experiment 2, primes were phonologically identical to the targets (homophones) but differed in their orthographic similarity. For LVF targets, only orthographic priming was significant, whereas for RVF targets, the phonological primes were effective regardless of their orthographic similarity. The results imply that orthographic activation, though maintained by both hemispheres, is more characteristic of right hemisphere word recognition processes, whereas phonological priming is more characteristic of left hemisphere processes.     

Orthographic and phonological priming in the two cerebral hemispheres

The patterns of activation invoked in the two cerebral hemispheres by written words may be different. Two lexical decision experiments investigated several aspects of such activation patterns. Experiment 1 tested phonological and orthographic priming in the hemispheres, manipulating two levels of phonological and two levels of orthographic similarity. Orthographic priming in the left visual field (LVF) was significantly larger than in the right visual field (RVF). In Experiment 2, primes were phonologically identical to the targets (homophones) but differed in their orthographic similarity. For LVF targets, only orthographic priming was significant, whereas for RVF targets, the phonological primes were effective regardless of their orthographic similarity. The results imply that orthographic activation, though maintained by both hemispheres, is more characteristic of right hemisphere word recognition processes, whereas phonological priming is more characteristic of left hemisphere processes.     

Finding the right word: hemispheric asymmetries in the use of sentence context information

The cerebral hemispheres have been shown to be differentially sensitive to sentence-level information; in particular, it has been suggested that only the left hemisphere (LH) makes predictions about upcoming items, whereas the right (RH) processes words in a more integrative fashion. The current study used event-related potentials to jointly examine the effects of expectancy and sentential constraint on word processing. Expected and unexpected but plausible words matched for contextual fit were inserted into strongly and weakly constraining sentence frames and presented to the left and right visual fields (LVF and RVF). Consistent with the prediction/integration view, the P2 was sensitive to constraint: words in strongly constraining contexts elicited larger P2s than those in less predictive contexts, for RVF/LH presentation only. N400 responses for both VFs departed from the typical pattern of amplitudes graded by cloze probability. Expected endings in strongly and weakly constraining contexts were facilitated to a similar degree with RVF/LH presentation, and expected endings in weakly constraining contexts were not facilitated compared to unexpected endings in those contexts for LVF/RH presentation. These data suggest that responses seen for central presentation reflect contributions from both hemispheres. Finally, a late positivity, larger for unexpected endings in strongly constraining contexts, observed for these stimuli with central presentation was not seen here for either VF. Thus, some phenomena observed with central presentation may be an emergent property of mechanisms that require interhemispheric cooperation. These data highlight the importance of understanding hemispheric asymmetries and their implications for normal language processing.     

N170 ERPs could represent a logographic processing strategy in visual word recognition

Background  

Occipito-temporal N170 component represents the first step where face, object and word processing are discriminated along the ventral stream of the brain. N170 leftward asymmetry observed during reading has been often associated to prelexical orthographic visual word form activation. However, some studies reported a lexical frequency effect for this component particularly during word repetition that appears in contradiction with this prelexical orthographic step. Here, we tested the hypothesis that under word repetition condition, discrimination between words would be operated on visual rather than orthographic basis. In this case, N170 activity may correspond to a logographic processing where a word is processed as a whole.     

A tale of two recognition systems: implications of the fusiform face area and the visual word form area for lateralized object recognition models

Two areas of current intense interest in the neuroimaging literature are that of the visual word form area (VWFA) and of the fusiform face area (FFA) and their roles in word and face perception, respectively. These two areas are of particular relevance to laterality research because visual word identification and face identification have long been shown to be especially lateralized to the left hemisphere and the right hemisphere, respectively. This review therefore seeks to evaluate their significance for the broader understanding of lateralization of object recognition. A multi-level model of lateralized object recognition is proposed based on a combination of behavioral and neuroimaging findings. Rather than seek to characterize hemispheric asymmetries according to a single principle (e.g., serial-parallel), it is suggested that current observations can be understood in terms of three asymmetric levels of processing, using the framework of the Janus model of hemispheric function. It is suggested that the left hemisphere represents features using an abstract-category code whereas the RH utilizes a specific-exemplar code. The relationships between these features are also coded asymmetrically, with the LH relying on associative co-occurrence values and the RH relying on spatial metrics. Finally, the LH controlled selection system focuses on isolating features and the RH focuses on conjoining features. It is suggested that each hemisphere utilizes efficient (apparently parallel) processing when stimuli are congruent with its preferred processing style and inefficient (apparently serial) processing when they are not, resulting in the typical left-lateralization for orthographic analysis and right-lateralization for face analysis.     

The role of the right cerebral hemisphere in processing novel metaphoric expressions taken from poetry: a divided visual field study

Previous research suggests that the right hemisphere (RH) may contribute uniquely to the processing of metaphoric language. However, most studies have focused on familiar metaphoric expressions. The present study used the divided visual field paradigm to examine the role of the right cerebral hemisphere in processing novel metaphoric expressions taken from poetry. In two experiments, participants were presented with four types of word pairs, literal, conventional metaphoric and novel metaphoric expressions and unrelated word pairs, and asked to perform a semantic judgment task. Two different SOAs: 400 and 1100 ms were used. The results of both experiments showed that responses to LVF/RH presented target words were more accurate and faster than responses to RVF/LH target words for novel metaphoric expressions, but not for other types of word pairs. These results support previous research indicating that during word recognition, the RH activates a broader range of related meanings than the LH, including novel, nonsalient meanings. The findings thus suggest that the RH may be critically involved in at least one important component of novel metaphor comprehension, i.e., the integration of the individual meanings of two seemingly unrelated concepts into a meaningful metaphoric expression.     

Do activated letters influence lexical selection in written word production?

Background: Written communication has become increasingly important since the advent of digital media such as emailing and texting. Research into the nature of interaction between the cognitive processes underlying production have focused on spoken language. Less is known about the processes underlying written word production.

Aims: We investigated the presence of feedback from letter activation to lexical selection in written word production. The presence of feedback would lead to greater-than-chance orthographic overlap between targets and semantically-related word errors in acquired dysgraphia, as the semantic neighbours of a target word sharing orthographic structure would receive extra support during lexical selection.

Methods & Procedures: The orthographic overlap of written semantically-related word errors from an individual with acquired dysgraphia was compared to chance distributions of overlap generated using a word association database. We used a Monte Carlo procedure to select random “pseudo-errors” from the top semantic associates for each target word. 10,000 hypothetical datasets were generated for two distinct hypotheses regarding the likelihood of selecting particular semantically-related word errors.

Outcomes & Results: The orthographic overlap between target words and semantically-related word errors for the individual with dysgraphia was greater than any of the hypothetical datasets, indicating that the target-error pairs shared more orthographic structure than expected by chance (p < .0001).

Conclusions: We obtained evidence that feedback from letter activation influenced lexical selection in a case of impairment. The presence of feedback from letter activation to lexical selection may be useful in developing strategies leading to improved language production in individuals with dysgraphia and aphasia.     

Perceptual discrimination of Shona lexical tones and low-pass filtered speech by left and right hemisphere damaged patients

Background: While the role of the right hemisphere (RH) in prosodic processing is prominent, research on the perception of lexical tones has shown that left hemisphere damaged (LHD) patients are more impaired than right hemisphere damaged (RHD) patients. Dichotic listening and imaging studies with healthy speakers of tone languages demonstrate that at least at the phonemic and lexical level, prosody is processed in the left hemisphere (LH) when the variations in pitch are phonemically distinctive. There is no report available yet on the perceptual discrimination of a Bantu language in patients after unilateral brain damage.

Aims: We addressed the question of how well Shona aphasic patients and right hemisphere damaged patients perceive pitch contrasts in Shona lexical words and also in their homologous low-pass filtered counterparts. We also sought to discover the validity of the current hypotheses on hemispheric lateralisation particularly the hypothesis on hemispheric lateralisation based on language function to account for the Shona data.

Methods & Procedures: A total of 7 LHD and 7 RHD patients and 14 healthy controls participated in two discrimination tasks that examined perception of lexical tone in (a) bisyllabic Shona words and (b) low-pass filtered stimuli. In both tasks the participants were tasked with judging the pitch as the same or different in 120 bisyllabic words and 120 low-pass filtered stimuli.

Outcomes & Results: The results demonstrated that the tonal discrimination of the LHD group was more reduced in comparison to the RHD group and control participants. However, the performance of the RHD patients was not error free relative to the control participants, although significantly better than the LHD patients in both tasks.

Conclusions: At least for the phonemic and lexical levels, brain damage to the dominant hemisphere results in lexical tone impairment for LHD patients, and cognitive load processing results in a subdued but good performance for RHD patients. The LH is therefore dominant for processing tone when it is lexically distinctive.     

Hemispheric Asymmetry for Word Naming: Effects of Frequency and Regularity of Pronunciation

Observers named visually presented words as quickly as possible when the words were presented to the centre of a viewing screen (Experiments 1, 5, and 6) or were flashed to either the left visual field/right hemisphere (LVF/RH) or right visual field/left hemisphere (RVF/LH) on each trial (Experiments, 2, 3, and 4). Words varied in frequency of occurrence in the language and in regularity of pronunciation. On lateralised trials, there was a RVF/LH advantage for both error rate and reaction time, with this hemispheric asymmetry for word naming being independent of both frequency and regularity. The RVF/LH advantage was also independent of whether the letters within a word were arranged horizontally or vertically. Error rates and reaction times were lower for high-than for lowfrequency words and lower for words with a regular pronunciation than for exception words. In addition, on centre trials the effects of regularity were larger for low-frequency words than for high-frequency words. However, when words were presented in the visual periphery, the effects of regularity were of the same magnitude for high-and low-frequency words. In view of the theoretical importance of the Frequency by Regularity interaction, this latter result suggests that word processing is qualitatively different in the visual periphery than in the centre of vision; indicating that perceptual asymmetries in a typical visual halffield experiment may be limited in what they can tell us about the relative contributions of the two hemispheres to processing words in the centre of vision.     

Sentence priming effects in the two cerebral hemispheres: influences of lexical relatedness,word order,and sentence anomaly

The present study examined lexical, syntactic and semantic message-level contributions to sentence priming in the left (LH) and right (RH) hemispheres. Participants made lexical decisions to laterally presented target words preceded by congruent, incongruent, syntactic, random and neutral incomplete sentences that either contained or did not contain a lexical associate of the target. In the LH, congruent associated sentences facilitated, and incongruent associated sentences inhibited, target word recognition extensively. A similar pattern, but attenuated, was found for the RH. In both the LH and RH, random associated sentences benefited word target recognition very moderately. Syntactic associated sentence primes that contain word-level information embedded in a meaningless message provided the condition in which the RH was benefited but the LH was not. The asymmetry of sentence type context effects in the LH and RH suggests some important differences in message-level processes available to each hemisphere.     

Variability of right hemisphere activation during semantic word processing in aphasic patients: an electrophysiologic study in three patients

After stroke, the interhemispheric reorganisation of the neural network implicated in language is hypothesized to be a function not only of the site of lesion but also of the residual impairment. With a multiple case approach, we tested this hypothesis in three chronic aphasic patients. Two patients, GE (capsulo-lenticular stroke) and JHN (fronto-temporal stroke) showed formal residual semantic difficulties, while the third patient (EG, large sylvian lesion) did not. Brain electric activity was analysed during a categorisation task of tachistoscopically presented words in the left and the right visual field. The temporal analysis of brain activity showed that both patients with semantic residual difficulties activated the right hemisphere (RH) during some steps of word processing. In the third patient, without semantic impairment, the RH was activated only during a short time period. Further more, RH activation was shown to be dependent on the visual field of word presentation. Phonological impairment was not predictive of RH activation. These results suggest that RH activation, particularly anterior regions, can occur during semantic processing of words as a function of semantic residual impairment.     

Evaluating a split fovea model of visual word recognition: effects of case alternation in the two visual fields and in the left and right halves of words presented at the fovea

Two experiments are reported exploring the effect of cAsE aLtErNaTiOn on lexical decisions to words and nonwords presented laterally or centrally. In line with previous research, Experiment 1 found that case alternation slowed lexical decision responses to words more in the right visual field (RVF) than in the left visual field (LVF). In Experiment 2, the words and nonwords were all presented centrally. There were three conditions, a condition in which the word and nonwords were presented in lower case letters, a condition in which the letters to the left of the central fixation were case alternated (e.g., aMbItion, mOdLants) and a condition in which the letters to the right of fixation were case alternated (e.g., collApSe, pireNtOl). Alternating the case of letters to the right of fixation slowed lexical decision responses more than alternating letter case to the left of fixation. The results provide further support for a split fovea account of visual word recognition according to which those letters of a centrally-fixated word that fall to the left of fixation are processed initially by the right cerebral hemisphere while those letters that fall to the right of fixation are processed initially by the left cerebral hemisphere, with the characteristics of the left and right hemispheres being revealed in the processing of initial and final letters in centrally presented words.     

Inferring the Nature of Semantic Processes by Varying Priming Procedure: A Reply to Koivisto and Laine

Koivisto and Laine (2000, this issue, henceforth KL) describe two semantic priming experiments in support of a new theory of semantic processing across the cerebral hemispheres. According to the model, automatic meaning activation spreads only within the left hemisphere (LH), and the LH is needed to activate word meanings within the right hemisphere (RH). Next, a LH expectancy process focuses attention on relevant meanings and suppresses irrelevant meanings; as no such process is available to the RH, close and distant meanings remain activated there. If LH processing does not produce a coherent interpretation, then the RH integrates distant or unexpected meanings into the context. This is a provocative account, deserving careful scrutiny. KL report priming data using weakly related category members to support two premises of this theory: (1) when semantic priming occurs within the LVF/RH, this is attributable to postlexical semantic integration, and (2) when priming can be attributed to automatic meaning activation, it only occurs within the RVF/LH. I argue here that KL's data cannot support these claims.     

Neural correlates of the left-visual-field superiority in face perception appear at multiple stages of face processing

Studies in healthy individuals and split-brain patients have shown that the representation of facial information from the left visual field (LVF) is better than the representation of facial information from the right visual field (RVF). To investigate the neurophysiological basis of this LVF superiority in face perception, we recorded event-related potentials (ERPs) to centrally presented face stimuli in which relevant facial information is present bilaterally (B faces) or only in the left (L faces) or the right (R faces) visual field. Behavioral findings showed best performance for B faces and, in line with the LVF superiority, better performance for L than R faces. Evoked potentials to B, L, and R faces at 100- to 150-msec poststimulus showed no evidence of asymmetric transfer of information between the hemispheres at early stages of visual processing, suggesting that this factor is not responsible for the LVF superiority. Neural correlates of the LVF superiority, however, were manifested in a shorter latency of the face-specific N170 component to L than R faces and in a larger amplitude to L than R faces at 220-280 and 400-600 msec over both hemispheres. These ERP amplitude differences between L and R faces covaried across subjects with the extent to which the face-specific N170 component was larger over the right than the left hemisphere. We conclude that the two hemispheres exchange information symmetrically at early stages of face processing and together generate a shared facial representation, which is better when facial information is directly presented to the right hemisphere (RH; L faces) than to the left hemisphere (LH; R faces) and best when both hemispheres receive facial information (B faces).     

Interhemispheric connectivity during lateralized lexical decision

The well‐established right visual field (RVF‐lh) advantage in word recognition is commonly attributed to the typical left hemisphere dominance in language; words presented to the LVF‐rh are processed less efficiently due to the need for transcallosal transfer from the right to left hemisphere. The exact stage for this hemispheric transfer is currently unsettled. Some studies suggest that transfer occurs at very early stages between primary visual regions, whereas other studies suggest that transfer occurs between the left visual word form area and its right hemisphere homolog. This study explores these conflicting accounts and finds evidence for both. Participants conducted a lateralized lexical decision task with both unilateral and bilateral display conditions. Connectivity analyses were conducted from magnetoencephalography signals that were localized to the left middle occipital gyrus (LMOG), right middle occipital gyrus (RMOG), left visual word form area (LVWFA), and right visual word form area (RVWA). Results from unilateral trials showed asymmetrical interhemispheric connectivity from the RMOG to LMOG and symmetrical interhemispheric connectivity between the LVWFA and RVWFA. Furthermore, bilateral presentations led to reduced interhemispheric connectivity between both homologous region of interest pairs. Together, these results suggest that lateralized word recognition involves multiple stages of interhemispheric interactions and that these interactions are reduced with bilateral displays.     

Evidence of interhemispheric transmission in laterality effects

The study was aimed at testing various models that can explain visual lateral asymmetries due to hemispheric specialization. In Experiments 1-3 the subjects had to perform a lateralized "go-no go" discrimination of words (primary task) either alone or in association with secondary tasks that interfered with the processing of the left hemisphere (ordered tapping) or the right hemisphere (finger flexion). In Experiment 4 the primary task was one of lateralized "go-no go" discrimination of faces while the secondary tasks were again those of ordered tapping and finger flexion. The results showed that in the case of word discrimination the advantage in speed of response in favour of the right visual field/left hemisphere (RVF/LH), which was observed for the primary task alone, did not change when the secondary task was added. This held true irrespective of whether the secondary task loaded the left or right hemisphere. The advantage for the left visual field/right hemisphere (LVF/RH) observed for face discrimination alone, disappeared when the secondary task interfered with the processing of the right hemisphere and did not change when the secondary task concerned the left hemisphere. It was concluded that each hemisphere is able to elaborate in parallel the incoming information, but, in normal conditions, interhemispheric transmission is responsible for the lateral asymmetries in perception (conditional interhemispheric transmission model).