The present data reveal significant vulnerability of working memory performance to intervening distraction in healthy volunteers. Specifically, response was slowed during the probe period of a delayed response task, when a face distractor was presented during the delay relative to when a scrambled distractor was presented. This distractor-related slowing was absent in Parkinson’s disease patients OFF medication, leading to relatively faster responding after distraction than in controls. Thus, Parkinson’s disease patients exhibited enhanced distractor resistance when they were OFF their medication. Slowing was reinstated by dopaminergic medication, as evidenced by the finding that responding of the same patients in the ON medication state did not differ from that of controls, although the direct comparison between the ON and OFF medication states did not reach significance. The improved performance on the delayed response task of patients OFF medication relative to controls contrasts with their impairment on the backward digit span test. The finding suggests that Parkinson’s disease can enhance or impair working memory performance depending on task demands. Specifically, in Parkinson’s disease patients, working memory representations were more resistant, not only to distraction, leading to improvement on the delayed response task, but also to backward reordering, leading to impairment on the backward digit span test.
The finding that Parkinson’s disease patients exhibited enhanced distractor resistance significantly refines our understanding of the mechanisms of cognitive deficits in Parkinson’s disease. These cognitive deficits have often been argued to resemble the consequences of frontal lesions (Lees and Smith, 1983
; Brown and Marsden, 1988
; Owen et al.
; Dubois and Pillon, 1997
: but see Owen et al.
), possibly reflecting disruption of striatal output (Owen et al.
). However, studies with lateral frontal lesion patients have revealed disruption rather than facilitation of distractor resistance during working memory (Malmo, 1942
; Chao and Knight, 1995
). Thus, in contrast, the current data suggest that some types of frontal function might be enhanced even beyond normal function rather than reduced in mild Parkinson’s disease.
The enhancement might reflect deficient dopamine levels in the striatum, and/or upregulated dopamine levels in the prefrontal cortex (Rakshi et al.
). A functional division has been proposed to exist between dopamine-dependent striatal updating processes and dopamine-dependent prefrontal maintenance processes (Bilder et al.
). Poor striatum-dependent updating might confer benefits in terms of distractor resistance, due to reduced responsiveness to new input, and a striatal locus of modulation is suggested by recent imaging data showing transient under-activation of the striatum during set shifting and updating in working memory (an n
-back task) (Monchi et al.
; Marklund et al.
). It might be noted that our data are less consistent with other recent theorizing, positing an important role for the striatum in dopamine-induced increases in filtering and distractor resistance in working memory (Gruber et al.
; McNab and Klingberg, 2008
Alternatively, the current finding might reflect indirect modulation of prefrontal cortex output, rather than striatal output, given recent models of dopamine function in the prefrontal cortex (Seamans and Yang, 2004
; Durstewitz and Seamans, 2008
) and known neurochemical reciprocity between dopamine in the prefrontal cortex and dopamine in the striatum. According to these models, dopamine facilitates the stabilization of currently relevant representation in the face of intervening distractors by directly acting at the level of the prefrontal cortex. The hypothesis that the enhanced distractor resistance reflects modulation of the prefrontal cortex is further strengthened by evidence from our recent pharmacological functional MRI study, in which the same task was employed. In this study distractor-related neural activity in healthy young volunteers was modulated by dopaminergic drugs only in the prefrontal cortex, and not in the striatum (Cools et al.
Our finding that Parkinson’s disease patients exhibit normal delayed response performance in the absence of significant distraction is consistent with previous reports, which have also shown intact performance on delayed response tasks that do not require complex processing (Fournet et al.
; Ketcham et al.
; Lewis et al.
; Fern-Pollak et al.
; for review see Cools, 2006
). On the other hand, the present finding differs from those described in some other recent studies of working memory deficits in Parkinson’s disease. For example, Moustafa et al.
) recently reported that Parkinson’s disease patients were impaired when ignoring distractors during the delay of an adapted version of the AX-CPT (continuous performance task). However, this deficit was found only in patients ON medication and not in patients OFF medication. An important difference between the present study and that previous study is that the previous effect was obtained in a shifting phase of the task, when subjects had to shift attention (and responding) away from previously relevant stimuli (i.e. the current distractors) towards newly relevant stimuli. The present data suggest that the deficit in Parkinson’s disease patients ON medication, but not OFF medication, observed in that study might reflect a combination of both a Parkinson’s disease-related shifting impairment as well as a selective enhancement of distractor resistance in patients OFF medication.
Significant distractor costs in healthy controls were observed only for face trials, and not for scene trials. This observation replicates previous findings (Yoon et al.
) and might reflect disproportionate distractibility by biologically salient stimuli. Although we refrain from emphasizing effects of disease or medication on the scene distractor costs, which are difficult to interpret, we report for completeness that such effects were not significant (both P
’s > 0.2).
The observation that distractor resistance was affected only in terms of reaction times, and not in terms of accuracy, probably reflects a lack of sensitivity of the current task to detecting error distractor costs. Indeed the distractor cost surfaced only in terms of reaction times and not in terms of errors even in healthy controls. The finding that Parkinson’s disease patients OFF medication nevertheless exhibited enhanced distractor resistance, if only in terms of reaction times, suggests that the disease potentiated the robustness or strength of current task-relevant representations. Future studies should employ tasks that are sensitive to error distractor costs to investigate whether Parkinson’s disease also prevents the disruption of these representations qualitatively, which should lead to higher relative accuracy as well as higher relative reaction times after distraction. Such a more sensitive paradigm might also be more adequate for definitively testing the hypothesis that distractor vulnerability in Parkinson’s disease is sensitive to restoration by dopaminergic medication.
The current study was designed to assess not only distractor costs but also switch costs. Previous results indicate that healthy subjects make more errors at probe when they had to switch attention between faces and scenes than when they attended to the same stimulus category on two consecutive (non-switch) trials (Cools et al.
). In the current study no such switch costs were obtained. On hindsight, this is perhaps not surprising, because we increased the duration of the encoding period from 1000 ms to 3000 ms, in order to prevent presumed limits on general cognitive speed in the patients. We argue that the short duration of 1000 ms was in fact essential for these error switch costs to surface at probe, presumably because the error costs reflected incomplete reconfiguration of attentional set. The duration of 3000 ms in the current study must simply have been long enough for all subjects to complete this reconfiguration process.
In conclusion, the present data demonstrate enhanced resistance to distraction in Parkinson’s disease patients, but only when they are OFF their medication. Thus, mild Parkinson’s disease is accompanied not only by cognitive inflexibility, as evidenced by impaired backward digit span as well as deficient set shifting and task switching observed in many previous studies (Cools et al.
), but also by aberrant cognitive stability in the face of distraction. We hypothesize that this pattern reflects deficient dopamine levels in the striatum, and/or upregulated dopamine levels in the prefrontal cortex. Preliminary support was obtained for this enhanced distractor resistance to be sensitive to restoration by dopaminergic medication, perhaps reflecting a restoration of the balance between dopamine in the striatum and dopamine in the prefrontal cortex.