فعال سازی معنایی در بیماران مبتلا به بیماری پارکینسون و خاموش لوودوپا

Research suggests that dopamine may exert a neuromodulatory influence on automatic spreading activation within semantic networks. In order to investigate the influence of dopamine depletion on semantic activation in Parkinson's disease (PD), nine patients with PD performed a lexical decision task when on and off levodopa medication. Eleven healthy controls matched to the PD patients in terms of sex, age and education also participated in the study. Both directly related word pairs (e.g., tiger – stripe) and indirectly related word pairs (word pairs related via a mediating word, e.g., chalk – black) were used to measure semantic activation across stimulus onset asynchronies (SOAs) of 270 msec, 520 msec and 1020 msec. Analysis of variance statistics revealed that the activation of directly related and indirectly related targets was slower for the PD group relative to the control group. Within group comparisons revealed further changes to semantic activation in PD patients off medication, with no activation of directly or indirectly related target words evident in PD patients off medication. These results further clarify the nature of dopamine's neuromodulatory influence on semantic activation, and suggest that the nature of altered semantic activation in PD may depend on the magnitude of dopamine depletion.

Parkinson's disease (PD) is characterized by degeneration of the nigrostriatal dopaminergic system (Mink, 1996) and in addition to movement disorders, subtle language processing deficits are a frequently reported feature of the disease. There is currently substantial evidence to suggest that even in the absence of overt dementia, many patients with PD exhibit deficits in lexical-semantic processing (Bayles et al., 1993, Gurd, 1996, Gurd, 2000, Lewis et al., 1998, Portin et al., 2000 and Randolph et al., 1993). While the results of emerging research have indicated that semantic processing deficits in PD may be related to alterations in dopaminergic transmission (Grossman et al., 2002, Watters and Patel, 1999 and Watters and Patel, 2002), the precise nature of dopamine's influence on semantic processing currently remains unclear. The objective of the present study was to use online semantic priming tasks to delineate the impact of dopamine depletion in PD on semantic activation. Semantic priming refers to the faster recognition of a target word when it is preceded by a related prime word (tiger – stripe) compared to an unrelated word (table - stripe). These semantic priming effects have often been attributed to automatic spreading activation (Collins and Loftus, 1975, Neely, 1977 and Posner and Snyder, 1975). Spreading activation theories are based on the assumption that concepts form an interconnected semantic network, with semantically/associatively related concepts stored closer together. It is thought that the processing of a prime word induces a temporary spreading of activation that lowers the activation thresholds of related target words. Thus, by varying the stimulus onset asynchrony (SOA) between presentation of a prime and target, it is possible to measure the temporal window over which activation occurs. Spreading activation theories also assume that activation dissipates with increasing distance within the semantic network. Thus, while activation of directly related targets (e.g., tiger – stripe) will typically occur in a semantic priming task, activation may also, to a lesser extent, spread to target words that are only indirectly related to the prime (e.g., lion – stripe, are only related via the mediating word ‘tiger’). Indeed, researchers have revealed significant semantic priming effects for both direct and indirect semantic relations in healthy adults (Angwin et al., 2004, Arnott et al., 2003, Hill et al., 2002 and Weisbrod et al., 1999), with some researchers also illustrating faster lexical decisions to directly related compared to indirectly related targets (Hill et al., 2002 and Weisbrod et al., 1999). Accordingly, manipulating both SOA and relatedness (direct/indirect) in semantic priming tasks may represent a comprehensive method of identifying the manner in which dopamine depletion in PD influences semantic processing. Certainly, the potential neuromodulatory influence of dopamine on semantic processing is already well recognized. Kischka et al. (1996) measured direct and indirect semantic priming effects in healthy participants who had ingested either a levodopa or placebo capsule. The results revealed reduced indirect semantic priming in the participants who ingested levodopa, which the researchers suggested was consistent with a dopamine-induced focusing of activation within semantic networks. Similarly, Copland et al. (2003) found that the ingestion of levodopa by healthy participants caused a focusing of activation and a dampening of weaker associations. In contrast, in another investigation of semantic priming in healthy participants who ingested levodopa, the results were not consistent with a focusing of activation (Angwin et al., 2004). Instead, the results suggested that the onset and decay of activation to both directly and indirectly related targets was occurring more quickly for participants who ingested levodopa. These results suggested that dopamine may also be capable of altering the speed of activation and decay within semantic networks. Specifically, Angwin et al. (2004) suggested that by increasing the signal-to-noise ratio of information processing, dopamine may speed processing of the prime word, leading to a faster onset and decay of semantic activation rather than a focusing of activation. Such findings suggest that a hyperdopaminergic state will result in the presence of both direct and indirect priming at short SOAs, but the absence of such priming effects at longer SOAs. Based on this semantic priming research, there is mounting evidence to suggest that dopamine may exert a neuromodulatory influence on semantic activation. Thus, it may be expected that striatal dopamine depletion in PD will result in a pattern of semantic activation opposite to that observed in healthy participants on levodopa. Specifically, dopamine depletion may be expected to slow the spread of activation to both direct and indirect semantic relations, such that direct and indirect priming effects are only evident at longer SOAs. Alternatively, dopamine depletion may also be expected to lead to unfocused activation within semantic networks, such that spreading activation to indirectly related concepts will be increased. Some researchers have observed delayed semantic activation in patients with PD (Angwin et al., 2005; Arnott et al., 2001), providing support for the influence of dopamine on the speed of semantic activation. More importantly, research has also demonstrated that such delays do not occur in an all-or-none manner, but occur along a continuum. For instance, Grossman et al. (2002) found a normal pattern of semantic activation in a subgroup of PD participants with intact sentence comprehension skills, whilst delayed semantic activation was evident in a subgroup of participants with poor comprehension skills. In contrast, Angwin et al. (2007) found delayed semantic activation both in PD patients with good and poor sentence comprehension skills, but the magnitude of the delay was larger in those patients with poor comprehension skills. Grossman et al. (2002) has suggested that the magnitude of cognitive slowing in PD, which may be manifest as delayed lexical retrieval, could be dependent on the extent of disruption to the dopamine dependent frontal–striatal circuitry. Accordingly, Grossman et al. suggested that delays in semantic activation may only be evident in PD patients with more substantial levels of dopamine depletion. This suggestion is also consistent with other research that has suggested that the striatum plays a key role in the modulation of information processing speed (Harrington et al., 1998, Poldrack et al., 2001 and Schubotz et al., 2000). Since the extent of dopamine depletion may differ markedly across individuals with PD, comparisons of semantic priming in the same group of PD patients when on versus off dopaminergic supplementation may provide additional insight into the influence of dopamine on semantic processing. When on dopaminergic therapy such as levodopa, dopamine deficiency in PD should be temporarily replenished (but see Cools et al., 2003), whilst dopamine levels will decline when patients are off levodopa. To date, only a small number of studies have investigated semantic priming in PD patients both on and off levodopa. Murdoch et al. (2000) and Arnott et al. (2000) observed a negative priming effect (i.e., faster reaction times to unrelated than related target words) in PD patients with mild to moderate PD when off medication, which Arnott et al. (2000) interpreted within the framework of the center-surround theory of inhibition (Carr and Dagenbach, 1990). This theory postulates that when difficulty retrieving semantic information about a prime word is encountered, the activation of semantic concepts closely related to the prime may be inhibited in order to prevent them from blocking the retrieval attempt. Accordingly, Arnott et al. suggested that due to weakened activation of the prime and/or increased noise within semantic networks in PD patients off levodopa, concepts semantically related to the prime word become inhibited and a negative priming effect is obtained. Angwin et al. (2006) also investigated semantic priming in people with mild to moderate PD, and whilst they failed to observe negative priming effects in PD patients off levodopa, they did find that semantic priming was more susceptible to disruption when an unrelated word was presented between a related prime and the target word. Angwin et al. suggested that such results could also potentially be explained by weakened activation of the prime word, which makes semantic priming more susceptible to disruption. Taken together, therefore, the results of semantic priming studies in PD have suggested that dopamine depletion can induce both a slowing of automatic semantic activation as well as weaker or unfocussed activation within semantic networks. Importantly, weaker or unfocussed activation only appears to be evident under conditions of medication withdrawal in PD, suggesting that the nature of altered semantic activation in PD may differ as a function of the magnitude of dopamine depletion. Specifically, whilst dopamine depletion may initially lead to slowed activation, semantic activation may become progressively unfocussed as the magnitude of dopamine depletion increases. Whilst there is evidence to suggest that such unfocussed activation may lead to negative priming effects (Arnott et al., 2000 and Murdoch et al., 2000), it is still unclear whether unfocussed activation may also lead to increased spreading activation to distantly related concepts. Increased spreading activation in PD would be predicted based on previous findings of reduced indirect priming effects in healthy adults on levodopa (Kischka et al., 1996). The present study seeks to provide further insight into the dopaminergic changes to automatic semantic activation in PD, by combining measures of direct and indirect semantic activation across time with the testing of PD patients both on and off levodopa. Attempts to measure automatic semantic activation, however, are often confounded by the influence of conscious or attention-based processes. For instance, pre-lexical expectancies and/or post-lexical semantic matching strategies may influence semantic priming effects, particularly under experimental conditions where longer SOAs are used and/or when a high proportion of nonwords or related word pairs are used (Neely, 1991 and Neely et al., 1989). To facilitate automatic semantic activation and to reduce the influence of attentional confounds on semantic priming, the use of a pattern mask after the prime word will be implemented in the present study. By presenting a prime word only briefly (e.g., 80 msec) with a pattern mask presented (e.g., a random series of letters) after the prime, a participant's awareness of the prime may be lowered and the impact of conscious or attention-based processing on the lexical decision may be reduced. Recently, significant semantic priming effects have been successfully obtained with the use of masking techniques (Deacon et al., 2000 and Ruz et al., 2003). The use of a neutral priming condition will also be implemented in the present study. It has been well established that while automatic semantic priming results in facilitation (defined as significantly faster RTs to related target words relative to neutral target words), strategic or controlled processing can result in both facilitation as well as inhibition (defined as significantly slower RTs to unrelated target words relative to neutral target words) (Neely, 1977). Hence, the inclusion of a neutral prime to measure facilitation and inhibition is necessary in order to determine whether priming effects reflect automatic or controlled processes. The aim of the present research was to measure automatic semantic activation across time in PD patients (both on and off levodopa medication) and healthy controls using a semantic priming task with three SOAs (270 msec, 520 msec and 1020 msec) and four prime conditions (directly related, indirectly related, neutral and unrelated). Although strategic processing is typically observed at longer SOAs, the three SOAs from 270 msec to 1020 msec were chosen in order to explore the temporal aspects of semantic activation similar to that implemented in previous research (Angwin et al., 2005 and Angwin et al., 2007). Further, a masked prime word was utilised in order to minimise the potential influence of such strategic processes. For healthy controls subjects, it was predicted that activation of directly and indirectly related concepts would be evident at both the 270 msec and 520 msec SOAs. It was further predicted that no activation would be evident at 1020 msec SOA for the healthy control group, consistent with the decay in automatic semantic activation for healthy adults (Stern et al., 1991). In contrast, it was hypothesized that dopamine depletion would result in alterations to automatic semantic activation for the PD group. For the PD group on medication, it was predicted that the activation of directly and indirectly related concepts would not emerge until the 520 msec SOA, and that this activation would persist across the 1020 msec SOA, consistent with suggestions of a delayed time course of semantic activation in PD (Angwin et al., 2007, Arnott et al., 2001 and Grossman et al., 2002). For PD patients tested whilst off medication, it was predicted that the increased magnitude of dopamine depletion would lead to further disruptions to semantic priming due to an unfocussed or weaker activation of the prime word. Specifically, it was predicted that testing PD patients off levodopa would result in either increased activation of indirectly related concepts (as evidenced by an increased magnitude of indirect priming) or that negative priming effects would emerge.

The current study investigated automatic semantic activation in PD patients on and off levodopa and healthy adults. Although the results of the present study appeared to support a neuromodulatory influence of dopamine on semantic activation, the results also suggested that the precise nature of this influence varies in PD, depending on the extent of dopamine depletion. Specifically, it is proposed that striatal dopamine depletion in PD initially leads to a slower time course of semantic activation and decay, but that increased levels of dopamine depletion may lead to unfocussed or weaker activation of prime words and, subsequently, to additional disruptions in semantic activation. One caveat of the present study that must be considered is that only a small number of PD patients were able to participate in the research whilst off levodopa. Therefore, the results of this research need to be validated using a larger cohort of PD patients.