Language processing is dynamic and requires the participation of both cerebral hemispheres. The left hemisphere (LH) is considered to be the language dominant hemisphere, however, the right hemisphere (RH) is also accepted to play an important role in language processing. The RH has been linked with processing of discourse, comprehension of inferences, ambiguity and metaphoric language, and underlying much of this, is its role in lexical-semantic processing [see for a review of RH language processing [1
]]. According to dynamic models of cognitive functioning bilateral lexical-semantic processing will involve both interhemispheric activation and inhibition [2
]. The language dominant LH is suggested to inhibit aspects of RH participation in order to maximize the efficiency of word level processing and meaning selection [2
]. Interhemispheric inhibition has been suggested to limit the RH's ability to perform to its maximum semantic processing capacity under normal processing conditions, and to therefore have led to an underestimation of the RH's semantic processing ability [5
Conditions that overload the LH's control mechanisms and facilitate the RH's release from inhibition are suggested to allow the RH to exceed its traditional performance on word level lexical-semantic tasks [6
]. There is, therefore, the potential for RH disinhibition to play a role in the increases in RH activity observed in neuroimaging investigations of neural reorganization [7
] and improvements found for behavioral online RH semantic processing performance [13
] following LH lesions. However, investigations of hemispheric contributions to semantic processing following unilateral lesions have not directly measured the impact of RH disinhibition or residual LH suppression on the performance of each hemisphere. Therefore, the current study will reexamine hemispheric contributions to semantic processing following LH lesions under conditions designed to manipulate the degree of interhemispheric activation and inhibition.
Existing theoretical insights into hemispheric performance during semantic processing in healthy adults under normal processing conditions (i.e., without manipulation of RH disinhibition) suggest that both hemispheres are capable of processing lexical-semantic information, and that hemispheric differences relate to an interaction between the strength of relatedness of items and the time-course allowed for semantic access [16
]. This time-course hypothesis suggests that activation in the LH begins broadly and includes both strongly and weakly related items, followed quickly by the focusing of activation to include closely related items only. Alternatively, the RH is slow to initiate, but once activation is achieved, the RH will maintain that activation for both closely and distantly related items over time [16
]. Accordingly, the RH's role in semantic processing appears to be to maintain activation of multiple interpretations that can be called upon by the LH when complex meaning integration is required [17
Theories of hemispheric semantic activation are based predominately on findings of divided visual field (DVF) semantic priming investigations. Semantic priming refers to the facilitation effect wherein a target word is responded to more quickly and accurately when it is preceded by a related word (prime), compared to an unrelated one. Semantic priming is therefore a measure of the spread of activation in the semantic system. The DVF methodology is used in order to assess activation in each hemisphere separately, via the presentation of prime and/or target letter strings to the left visual field (LVF) or right visual field (RVF). Prime and target pairs related via category membership and/or association have been one focus for investigations into the time-course of hemispheric activation. These stimuli represent a hierarchy of activation that can differentiate hemispheric patterns of activation with respect to the interaction between relationship strength and time-course of semantic activation [e.g., [16
]]. Particular patterns of activation are suggested for each relationship type.
Items that are related via both category membership and association (e.g., DOG--CAT) are suggested to be the most strongly related pair type, compared to those related via category membership or association alone. This is due to the additive nature of the semantic (category membership) and lexical (association) relationships, and because these items are suggested to share the most semantic features [e.g., [19
]]. Evidence is accumulating that items related by category membership and association generally exhibit bilateral activation over time [18
Items related via association but not category membership (e.g., DOCTOR--HOSPITAL) are considered the next most strongly activated due to the benefit of the associative relationship [21
]. Behavioral and event-related potential priming studies have found predominately LH advantage for priming association only pairs, prior to 500 ms stimulus onset asynchrony (SOA; time between onset of prime and target) [22
]. There have been some suggestions that the RH is capable of priming this pair type at longer SOAs, in accordance with the time-course hypothesis [22
Finally, items that are related via category membership but not association (e.g., CAMEL--DOG) represent the weakest relationship. These items may share some semantic features but do not benefit from the associative boost, and often exhibit the weakest priming effect [21
]. Investigations of items related via category membership only suggest that the variation in hemispheric processing patterns relate to underlying processing conditions, with a LH asymmetry under automatic spreading activation conditions [25
] and a RH asymmetry under conditions that encourage strategic processing [26
] or at long SOAs [16
DVF priming investigations have provided insight into the semantic processing of the healthy brain under normal processing conditions. However, the RH disinhibition hypothesis suggests that RH performance described in such studies does not reflect its true capacity [5
]. Specifically, the RH disinhibition hypothesis suggests that the RH is less active and does not perform to capacity when normal LH dominance remains, however when this balance is modified the RH is able to demonstrate a greater degree of semantic processing e.g., [29
]. Supporting evidence for the RH disinhibition hypothesis has been derived from investigations of people with LH lesions [30
] and investigations that attempt to release the RH from LH suppression via experimental manipulations, such as short presentation times for lateralized stimuli [30
], incorporation of a secondary task to "overload" LH processing [5
], or simulation of lesions using rTMS [11
]. These investigations have found it possible to release the RH from unilateral suppression by manipulating the attentional balance between the hemispheres.
Manipulation of the interhemispheric attentional balance has been most specifically assessed using a combination of the dual task and DVF paradigms with healthy participants. The dual task paradigm takes advantage of functional task lateralization in order to either pre-activate or overload the attentional resources of a particular hemisphere. The concurrence decrement effect, that occurs when a particular hemisphere is overloaded, can be explained using the multiple limited resources model [32
]. The multiple limited resource model dictates that each hemisphere has a finite supply of attentional resources which cannot be shared. Therefore, when two sufficiently difficult concurrent tasks are performed by a particular hemisphere, the resource demands overlap to such an extent that the hemisphere becomes overloaded. Dual task investigations of semantic processing have used secondary verbal memory tasks in order to overload the LH's resources and produce a RH disinhibition effect and subsequent improvements in RH processing [e.g., [5
]]. The combination of DVF presentation in a dual task experiment allows the direct investigation of individual hemispheric performance following disinhibition.
DVF dual task investigations have found improvements in RH processing following disinhibition. Specifically these investigations have reported that RH disinhibition results in the reduction of the traditional LH processing advantage for lexical decision tasks, and RH activation of items traditionally associated with LH processing, such as long low imageable words [5
]. Similarly, participants with partial disconnection between the hemispheres have also exhibited an enhanced RH semantic processing performance following RH disinhibition under dual task conditions [29
The findings from DVF dual task investigations [e.g., [5
]], and in turn the RH disinhibition hypothesis, have interesting implications for the study of changes in hemispheric contributions to semantic processing for people who have suffered lesions to the language dominant LH. For instance, behavioral investigations of hemispheric contributions to semantic processing for people with LH lesions and subsequent aphasia, have found evidence of improved RH lexical-semantic abilities following LH lesions [13
] and neuroimaging investigations have found increases in RH activation during language tasks following LH lesions [7
]. Furthermore, investigations of spatial attention treatment for people with LH lesions have explored the effect of manipulating the participation of each hemisphere, and have found that priming or recruiting attentional or intentional mechanisms in the intact RH (via orienting attention to the hemispace or a complex left handed movement) can improve language function in some people with aphasia [33
]. However, the potential influence of RH disinhibition is yet to be fully explored in the context of the individual hemispheric contributions to language processing following LH lesions.
The current investigation seeks to extend the examination of changes to RH language function post LH lesion, in order to more directly explore the impact of the lesion on RH disinhibition and to explore the impact of RH disinhibition on RH performance. Therefore the current investigation will compare hemispheric performance on a DVF semantic priming task (primary task) over three verbal memory load conditions (concurrent task). Semantic priming effects may arise from one of three mechanisms; automatic spreading activation, controlled expectancy-based activation and controlled post-lexical semantic matching [36
]. The current investigation will encourage controlled activation by using a high relatedness proportion (ratio of related word targets and unrelated word targets; the higher the proportion the more controlled) [36
], as hemispheric differences appear to occur under controlled processing conditions [16
]. A shorter SOA (250 ms) was employed to allow post-lexical controlled processing, as this onset asynchrony is insufficient for the generation of expectancies. Post-lexical semantic matching occurs after the presentation of the target and before the lexical decision is made [36
] and involves checking from target back to prime, to assess semantic relatedness. The presence of a semantic relationship indicates that the target is a real word and therefore accelerates a word response [38
The stimuli conditions (category and association, category only and association only items) chosen for the current primary semantic priming task were included because of their sensitivity to hemispheric priming differences [e.g., [19
]]. In particular, the association only condition is hypothesized to provide a platform for the RH to exhibit an increase in priming performance, with previous findings suggesting that the RH exhibits some difficulty processing items related via association only in healthy adults at shorter SOAs [22
]. However, given that both hemispheres have access to the same semantic network, following disinhibition the RH will become more active and have the potential to activate all of the related conditions. It is also noted that association only stimuli have been found to prime during central priming investigations of people following LH lesion [e.g., [40
]], therefore it is plausible that the RH is capable of activating these stimuli under conditions that promote RH performance.
It is hypothesized that a different pattern of priming is expected for each of the related stimuli conditions depending on the degree to which RH disinhibition is occurring. The related conditions are expected, under normal processing conditions (ie. no disinhibition), to exhibit bilateral priming for items related by category membership and association, priming in the LH for the association only condition and priming in the RH for the category only condition, as well as exhibiting an overall LH response time advantage. Under conditions that elicit RH disinhibition it is expected that RH priming will increase to include significant priming of the association only condition, and that the overall LH processing advantage will be removed.
For control participants it is expected that the 6-word concurrent memory task will sufficiently modify the attentional balance between the hemispheres, and in accordance with the RH disinhibition hypothesis, result in an increase in RH priming. It is hypothesized that performance by participants with a LH lesion will differ from the control participants, as they will be more susceptible to the demands of the concurrent task, and to the impact of the lesion itself on hemispheric control. If the LH lesion results in disinhibition of RH semantic processing, then participants with a LH lesion should exhibit similar activation during both baseline and memory load conditions, including priming of all three related conditions in the RH. Conversely, if the LH lesion does not result in complete RH disinhibition, then improvement in RH processing is expected to improve under the 2- word and 6-word conditions, with priming of association only stimuli occurring in the RH when the concurrent load is sufficient to overload LH resources.