Effect of disease severity and dopaminergic medication on recollection and familiarity in patients with idiopathic nondementing Parkinson's

Abstract

The effect of disease severity and dopaminergic medication on the assessment of familiarity and the recollection of episodic details during recognition in nondementing idiopathic Parkinson's is uncertain. Some studies have reported familiarity as deficient in mild Parkinson's yet others have found it intact even in moderate Parkinson's. Recollection has been found to be both preserved and deficient in mild and moderate Parkinson's. The extent to which these conflicting findings are explained by disease severity or dopaminergic medication or a combination of the two is uncertain, as all studies assessed patients in a medicated state, and disease severity has not always been consistently reported.

Twelve patients with mild Parkinson's and 11 with moderate Parkinson's (medicated Hoehn and Yahr mean: 2.1 and 3.2, respectively), completed matched versions of a yes/no recognition memory test in a medicated and unmedicated condition (termed ON and OFF, respectively). Twenty-one matched healthy volunteers also completed both memory tasks in 2 separate sessions (termed Blue and Green, respectively).

In the ON/Green condition, the moderate Parkinson's recollection performance was significantly poorer than the healthy volunteers and mild Parkinson's. By contrast, recognition memory and familiarity measures in both Parkinson's group were relatively spared. In the OFF/Blue condition, the moderate Parkinson's recollection was impaired, but only in relation to the healthy volunteer set. There were no significant differences in recollection performance between the mild and moderate Parkinson's groups. Again, recognition memory and familiarity measures in both Parkinson's group were relatively spared. Further analyses showed the moderate patients’ recollection rates to be significantly poorer ON-medication compared to OFF.

These findings are discussed in relation to the staging of disease progression on medial temporal areas which separately support recollection and familiarity, and the putative effects the different classes of dopaminergic drugs may have on these areas.

Introduction

This study investigated the effect of dopaminergic medication and disease severity on the assessment of familiarity and the recollection of episodic details during recognition in patients with nondementing idiopathic Parkinson's. These processes differ with respect to the type of information that they provide and the level of recognition confidence each typically produces. A widely held view holds that recollection is a high confidence threshold process that involves remembering specific details from episodic memory regarding a past event. By contrast, recognition based on feelings of familiarity varies continuously as a reflection of memory strength in the absence of retrieval of contextual detail (Yonelinas & Jacoby, 1995).

There is conflicting evidence regarding the status of recollection and familiarity at different stages of Parkinson's, with only one study to date examining recollection in patients in mild and moderate stages of disease severity (Hay, Moscovitch, & Levine, 2002). In this study, recollection was normal in the mild Parkinson's group (Hoehn & Yahr, 1967 rating severity ratings in the range of 1–2.5) but significantly declined in the moderate group (Hoehn and Yahr [HY] 3–4). Familiarity was not assessed. Three other studies have investigated the dual process view of recognition memory, sampling patients at a single disease stage. It is important to note that there was consistency between these studies in relation to their use of the remember–know paradigm and adoption of the formulae provided by Yonelinas and Jacoby (1995) to derive estimates of recollection and familiarity. Consistent with Hay et al.’s findings, recollection was spared in mild Parkinson's (mean illness duration = 5.79 years [HY not provided], Davidson, Anaki, Saint-Cyr, Chow, & Moscovitch, 2006) and deficient in moderate Parkinson's (HY 2–3, Edelstyn, Mayes, Condon, Tunnicliffe, & Ellis, 2007), although Barnes, Boubert, Harris, Lee, and David (2003) reported sparing of recollection in moderate Parkinson's (mean HY 2.86) unless the patients had a history of visual hallucinations (mean HY 3.39).

The effect of disease severity on familiarity is also uncertain. In the study by Davidson et al. (2006), familiarity was impaired in mild Parkinson's, whereas both Barnes et al. (2003) and Edelstyn et al. (2007) found it to be preserved in moderate Parkinson's. We have identified four reasons (there may be others) why evidence is not fully concordant. First, although each of these studies assessed medicated patients, not all may have been in an optimally medicated state (e.g., Hay et al., 2002). Second, there is considerable variation between studies in the mode of classifying disease stage. Third, neuropsychological characteristics of patients varied, with executive dysfunction present in some patients (Barnes et al., 2003, Edelstyn et al., 2007, Hay et al., 2002) but not others (Davidson et al., 2006). Fourth, a major problem of studies in this area is the accurate measurement of recollection and familiarity, and particularly of familiarity. This problem applies most strongly to the remember/know procedure where it is well established that procedural differences, inadequate training of participants and inadequate attempts to ensure participants understand the procedure, as may be the case when the remember–know procedure is used in a surprise memory test (e.g., Barnes et al., 2003).

An influential view in Parkinson's research, is that deficits in free recall (Daum et al., 1995, Gabrieli et al., 1996, Ivory et al., 1999, Johnson et al., 2005) and recollection (Hay et al., 2002, Barnes et al., 2003) are contingent on a breakdown in prefrontally mediated memory processes underlying long-term memory encoding and retrieval strategies (such as the use of semantic organisation). Strategic memory processes are likely to depend, at least in part, on executive functions such as planning, decision-making and working memory (Shimamura, Janowsky, & Squire, 1991), which therefore places the origin of the recollection deficits within the mesostriatal-frontal system. However, there is a growing body of evidence suggesting disruption of dopamine modulation of mesolimbic structures, includes the ventral tegmental area and the hippocampus, may also contribute to recall and recollection impairments in Parkinson's. Evidence in support of this proposal is reviewed below.

Firstly, animal studies demonstrate a critical role for dopamine in inducing hippocampal long-term potentiation, a form of synaptic plasticity thought to underlie memory storage (Lemon and Manahan-Vaughan, 2006, Mockett et al., 2004, Otmakhova and Lisman, 1996, Wood et al., 2006), mediated by D₁, D₃, D₄ and D₅ dopamine receptors in the CA1–3 fields of the hippocampus (Bentivoglio & Morelli, 2005, chap. 1; Li et al., 2003, Mockett et al., 2007, O’Carroll et al., 2006). Dopamine has also been shown to modulate synaptic plasticity in the perirhinal cortex (Bentivoglio & Morelli, 2005, chap. 1; Cummings et al., 2006, MacDonald et al., 2009), mediated by D₂ receptor (Bentivoglio & Morelli, 2005, chap. 1). It should be noted at this point, that involvement of the hippocampus in recollection and the perirhinal cortex in familiarity has already been proposed in Aggleton and Brown, 1999, Aggleton and Brown, 2006 still controversial dual process model of episodic memory.

Secondly, anatomical evidence supporting a role of the mesolimbic circuit in memory comes from a series of functional magnetic imaging studies of healthy older adults. These investigations report a positive correlation between memory formation and integrity/activation of the ventral tegmental area (Bunzeck et al., 2007, Düzel et al., 2008, Kumaran and Düzel, 2008, Wittman et al., 2008). Whilst it appears that reward-related activation of the medial substantia nigra pars compacta is associated with improved hippocampus-dependent memory consolidation (Wittman et al., 2005), encoding-related midbrain activation also occurs independently of reward (Schott et al., 2004). Other evidence that genetic polymorphisms in the dopamine clearance pathways, such as the dopamine transporter (DAT1), affect encoding-related activation patterns in the midbrain and hippocampus (Schott et al., 2006) further supports the case that dopamine plays an important role in memory.

Thirdly, Braak et al. (2003) (see also Braak and Del Tredici, 2008, Braak et al., 2006) examined the brains of 41 patients obtained at autopsy, (clinical severity and cognitive function unknown at time of death). They proposed that α-synuclein pathology, the most abundant protein in Lewy bodies, spreads in a predictable caudo-rostral direction through the brain, beginning in the medulla oblongata and midbrain, before extending to the CA2 fields of the hippocampus and the transentorhinal region (i.e. the medial portion of the perirhinal cortex, BA 35/35a, Garey, 1999, Van Hoesen and Pandya, 1975, Taylor and Probst, 2008) and on to the primary sensory and association areas and prefrontal cortex (see Kalaitzakis, Graeber, Gentleman, & Pearce, 2008 for a critical discussion of this controversial model). Building on Braak et al.’s staging model, memory impairments are predicted based on medial temporal lobe pathology, and furthermore, hippocampal-dependent memory processes will decline prior to perirhinal-dependent processes, due to relative sparing of lateral perirhinal areas. However, the staging of these memory changes in relation to clinical severity is unclear.

Magnetic resonance (MR) imaging studies of Parkinson's patients using manual volumetric and voxel-based morphometry suggests recall impairments emerge, even in the mild–moderate stages (mean HY 2.5, Ibarretxe-Bilbao et al., 2008) once hippocampal pathology has reached a critical level (Brück et al., 2004, Camicoli et al., 2003, Camicioli et al., 1999, Duda et al., 2009, Laakso et al., 1996; but see Burton et al., 2004, Bouchard et al., 2008, Beyer et al., 2009, Dashtipour et al., 2009, Jokinen et al., 2009).

In sum, the findings from animal research and MR studies of healthy volunteers support a role for dopamine modulation of hippocampal and perirhinal memory processes, and furthermore, post-mortem studies of the brains of Parkinson's patients and MR studies of the hippocampus in nondementing Parkinson's patients indicate both structures are subject to the development of staged pathology.

Dopaminergic medication, and here we are primarily considering l-dopa, can also have a significant effect on cognitive function. Evidence indicates that the requisite dopaminergic state necessary to control motor symptoms has the potential to move the same patient away from their optimum for certain cognitive functions (see “l-dopa overdose hypothesis”, described by Gotham et al., 1988, Cools, 2006, Cools et al., 2001, Rowe et al., 2008), and may even lead to a dopamine dysregulation syndrome, marked by an increase in risk-taking behaviour such as pathological gambling and hypersexuality (Driver-Dunckley et al., 2003, Dodd et al., 2005). The relationship between the efficiency of neuronal activity and the state of dopaminergic modulation in the l-dopa overdose hypothesis is represented by a Yerkes-Dodson inverted U-shaped curve with cognitive functions declining with deviation away from optimum dopamine levels, indicated by the centre of the curve. Extrapolating this model to recollection and familiarity, implies that l-dopa has the capacity to both improve and impair these kinds of memory depending on baseline dopamine levels in the underlying neural circuitry.

The aim of our study was to investigate the impact of disease severity and dopaminergic medication on familiarity and recollection in nondementing idiopathic Parkinson's. The predictions for our study have been informed by the application of the l-dopa overdose hypothesis (Cools, 2006) on dopamine-dependent medial temporal lobe memory circuits, the staging model of α-synuclein pathology in Parkinson's (Braak et al., 2003) and the neural correlates of the dual process view of recognition memory (Aggleton and Brown, 1999, Yonelinas et al., 2002). Our central hypothesis was that the neural pathways on which recollection and familiarity separately depend, are differentially affected by disease progression, and consequently, the effects of dopaminergic medication on these memory measures will also differ.

The method used in this study to assess the effects of l-dopa on recognition memory, familiarity and recollection is the controlled l-dopa withdrawal procedure. This requires patients to abstain from their dopaminergic medication for a (“wash out”) period of 12–18 h prior to the memory assessment. Performance in this OFF state is compared with performance on a separate testing session, taking place at the same time of day as the OFF state, during which patients take their routine medication as usual. This procedure minimises, but does not eliminate, any effects that dopaminergic medication may have on recollection and familiarity. It is also less prone to the confounds of differences in disease severity compared to the alternative of comparing de novo, i.e. never medicated, patients with the same individuals at a later stage after l-dopa administration, or a different already-treated group.

Our first set of predictions applied to patients tested in an unmedicated state. In mild Parkinson's, we expected both recollection and familiarity to be preserved, contingent on the relative preservation of both the hippocampus, lateral and medial perirhinal cortical areas. By contrast, in moderate Parkinson's, a significant decline in recollection was predicted, contingent on developing pathology in the hippocampus. By contrast, relative sparing of familiarity performance was expected, due to preservation of lateral (if not medial) areas of the perirhinal cortex.

The second set of predictions pertained to the performance of the same patients assessed in a fully medicated state. We expected a l-dopa induced impairment of both recollection and familiarity in mild Parkinson's, as routine medication overdosed (close to) optimal dopamine levels in the hippocampus and perirhinal cortex. By contrast, in moderate Parkinson's, l-dopa should have a beneficial effect on recollection, as medication remediated depleted dopamine levels in the hippocampus. Familiarity performance was again expected to show relative sparing, as the same l-dopa which ‘overdosed’ optimal dopamine levels in the lateral perirhinal cortex also restored depleted levels in medial perirhinal cortex.