Dopaminergic modulation of cognitive interference after pharmacological washout in Parkinson's disease

Abstract

The dopaminergic modulation of prefrontal function in Parkinson's disease (PD) has been consistently demonstrated. There is evidence that the effects of pharmacological manipulations on cognitive performances are described by an “Inverted-U” shaped curve. Neuroimaging studies performed before and after an overnight withdrawal from therapy showed significant differences between drug states, but did not control for the relative impact of the long duration response to levodopa. Here we evaluate the brain response after a complete pharmacological washout by correlating dopaminergic-related changes of this response to changes in performance during cognitive interference. Twelve idiopathic PD patients were studied with functional MRI while performing a modified version of the Stroop task. Patients were scanned twice: (1) following a prolonged washout procedure (“OFF” state) and (2) 90–120 min after the administration of levodopa (“ON” state). Task-related changes of PD patients were compared to those of matched healthy controls. Healthy controls displayed prefrontal and parietal responses that were positively correlated with task accuracy. In the “OFF” state, PD patients showed significant responses in anterior cingulate and pre-supplementary motor area, which are hypothesized to operate at a higher level of basal dopaminergic modulation. Levodopa administration attenuated such responses and enhanced the response of prefrontal cortex (PFC), which was correlated with improved accuracy. Results demonstrate that the behavioral effects of pharmacological manipulations of the dopamine system are highly dependent on the baseline status of PFC. When a true hypodopaminergic state is induced in PD patients, cognitive interference might significantly benefit from the administration of levodopa via an enhanced PFC response.

Introduction

It is recognized that patients with Parkinson's disease (PD) show executive dysfunction [17], [34], [22], which may predict incident dementia [43]. Functional imaging studies invocated a role for disruption of both nigro-striatal [31], [10] and mesocortical dopaminergic pathways [28]. Although levodopa can improve high-level cognition [23], [15], [38], it can also have no effects, or sometimes selectively exacerbate cognitive deficits [37]. This does not depend on neuropsychological specificity of the drug, disease duration or severity. A high variability across cognitive tasks suggests that the magnitude and the direction of the effects of levodopa depend on cognitive demand [8], but they also reflect the amount of dopamine (DA) depletion and the level of DA receptors stimulation in the prefrontal cortex (PFC). The relationship between dopaminergic function and cognitive performance follows an “Inverted-U” shaped curve [1], suggesting the need to be in an optimal DA range. Thus, the manipulation of the DA system will be profoundly influenced by baseline individual DA levels in the PFC.

Functional MRI (fMRI) demonstrated that after administrating dextroamphetamine, healthy subjects show an “Inverted-U” shaped relationship between PFC function and working-memory. Dextroamphetamine induced a right shift of DA function in the PFC, as suggested by improved accuracy (optimal stimulation) in poor performers at baseline, but supraoptimal (detrimental) stimulation in good performers [26].

PD offers the opportunity to investigate the pharmacological manipulation of the DA system in humans. Neuroimaging studies evaluated the cognitive effects of a controlled levodopa withdrawal in PD patients, by using ON–OFF approaches [28], [9]. The administration of levodopa induced task-related changes in the PFC response, associated with better performance on specific executive tasks. A limitation of these studies is that they were based on a single overnight withdrawal procedure and therefore they did not account for the long duration response (LDR) to levodopa, a sustained motor benefit that builds up over days and likewise decreases over days after withdrawal [33]. Stable PD patients under chronic treatment are likely close to their maximal therapeutic gain, as they benefit from the LDR. The relatively hypodopaminergic state induced by an overnight drug withdrawal may partly mask the effects of levodopa on the DA system, because of the high variability in terms of medication status and disease progression [45]. Conservative protocols, designed to abolish the LDR, will target the suboptimal-optimal range of the “Inverted-U” curve and might be more sensitive in picking up differences between “ON” and “OFF” states.

Here we used blood oxygenation level dependent (BOLD) fMRI to test whether: (i) levodopa-induced changes in cognitive performance of PD patients in a complete baseline condition correlate with changes in the BOLD response; (ii) drug effects are still described by an “Inverted-U” curve within the task-related brain network.

The functional response was assessed with the Stroop task. This task relies on the interference of word reading upon color naming, when a color name is displayed using a different color, semantically incongruent. Several investigators using different versions of this task found activations in the anterior cingulate/paracingulate cortices (ACC/pre-supplementary motor area, pre-SMA), parietal cortex and lateral PFC [2], [6], [32], [40]. Conflict resolution is a cognitively demanding condition and is strongly expected to expose impairment in PD patients undergoing a pharmacological washout.