Dementia in Parkinson Disease: Current Concepts in Neuropathology, Neuroanatomy and Neurochemistry

From a clinical point of view, a better clarity regarding the nature of dementia in Parkinson disease and other parkinsonian disorders came with theories that two types of dementia could occur in neurodegenera- tive disorders. These were coined “subcortical” dementia and “cortical” dementia (Albert, Feldman, and Willis 1974; Cummings 1986). The entity “subcortical” dementia was associated with Parkinson disease and was characterized by problems with attention and deficits in spatial working memory and cognitive planning similar to those seen in animals with pre- frontal damage (Mishkin 1957; Shallice 1982). Descriptions of “subcortical dementia” were often ascribed to a slowness of thought or “bradyphrenia” that results in prolonged cognitive processing time most evident when tasks become more complex (Zimmerman et al. 1992; Cooper et al. 1994). “Cortical” dementia had Alzheimer disease as its prototype and centered on language abnormalities and the presence of apraxia and agnosia (Cum­mings 1986). While the terms “subcortical” and “cortical” dementia are now rarely used, they nonetheless have set a valuable framework to under­stand how cognitive dysfunction in Parkinson disease might be rationally organized.

The best present-day correlate to “subcortical” dementia is executive cognitive dysfunction. Executive cognitive function represents a battery of intellectual functions centered on attention span, concentra­tion, and memory. Inclusive in these functions is an assortment of compli­cated processes that are involved in decision-making such as anticipation, judgment, motivation, social/ethical appropriateness, and goal-directed behavior. A fundamental requirement of executive cognitive function involves the ability to store and hold information on-line so that it can be continuously accessed in order to plan and change strategies. These func­tions, as will be discussed later in this post, are believed to be mediated through dopaminergic pathways involving midbrain neurons in the sub­stantia nigra and ventral tegmental area that are either directly or indirectly distributed to the prefrontal cortex. The concept of executive cognitive function implies the requirement of networks to process information from multiple areas of brain as a necessary part of decision making.

At the core of this processing lies “working memory,” which involves prefrontal dopaminergic neurons that have the capacity to maintain firing as a probable method of storing information when tasks are being planned and executed (Goldman-Rakic 1996). One can envision initiating an activ­ity that requires a wide range of information to accomplish. For example, throwing an initial punch in a fight would not only require the gathering of visual, proprioceptive, and motor information but would also process more complicated data such as the ethics and social consequence of such an action. In addition, long-term memory information would be likely included (e.g., What happened when I previously performed this action?). The process of storing all information as it is gathered prior to actual out­put appears to be an important function of prefrontal neurons mediating working memory. The prefrontal area is ideally situated to receive infor­mation from the neocortex as well as the limbic system. A more complete gathering and maintaining of information in prefrontal areas prior to ini­tiating a motor or cognitive task could be viewed as a neurochemical/ neurophysiological correlate for the concept of logic. That is, all input has been considered for a task and consequently the output has been a decision based on the processing of the most complete database possible. Conversely, an inability to employ a full database because either the pre- frontal working memory cells cannot receive information from all parts of the brain or they cannot maintain a firing rate as information is gathered would result in an output that has not considered a full complement of potentially available information for planning. The result may be more impulsive appearing behavior that appears illogical.

One of the best tests capable of demonstrating executive cognitive dys­function in Parkinson disease and the role of “working memory” is the Wisconsin Card Sorting Test. In this test “stimulus cards” with shapes of different colors, designs, and quantities are initially presented to the sub­ject. Without informing the subject, the examiner then decides whether additional cards to be given to the subject will be matched to the stimu­lus cards by color, design, or quantity. The examiner tells the subject only whether the match of the new additional cards to the stimulus cards is right or wrong. Consequently, by trial and error, the subject learns how he or she should be matching the cards. The matching rules are subse­quently changed (without informing the subject) and the subject must change strategy to learn and apply new rules. The test measures how long it takes the subject to learn the correct strategy a well as how long it takes to acquire the new strategy (“change sets”). Consequently, the Wisconsin

Card Sorting Test gives information on how well the subject can plan and change strategies based on goal-directed behavior, all of which are fun­damental skills related to executive function. It provides information on both impulsive and perseverative tendencies. It is easy to envision how tasks associated with the Wisconsin Card Sorting Test would depend heavily on maintaining activity in working memory cells since decisions are closely based on information gathered in previous successful match­ing attempts. In early Parkinson disease patients, deficits are seen in the Wisconsin Card Sorting Test with a prominent feature being the tendency to perseverate; that is, patients find it difficult to learn new strategies when the task has changed (Canavan et al. 1989; Paolo et al. 1996). Clinically, behavioral perseveration is frequently manifested in some patients when they seem unable to change their actions even when such previous behav­ior has resulted in an unwanted outcome (e.g., persistent falling because of not using an available cane or walker, continuing to drive a car despite previous accidents). Others have also found problems in early Parkinson disease involving visual working memory with a sparing of verbal work­ing memory (Bradley, Welch, and Dick 1989). Of interest is the observation that there appears to be preservation of visual working memory for shapes (Postle, Jonides, and Smith 1997; Owen et al. 1997). It is suggested that the visual working memory deficits become more evident in early Parkinson disease only when more complex executive function tasks are added to the testing paradigms such as set shifting (Owen 2004). Abnormalities in the Ravens Progressive Matrices and Tower of London Test also suggest that the greatest deficits are evident when tasks are changed and new strate­gies need to be used (Farina et al. 2000; Owen, James, and Leigh 1992). Per­severation of previous strategies and inconsistent performance on newly acquired strategies tend to predominate.

The Ravens Progressive Matri­ces tests a subject’s ability to find patterns in an apparent chaos of visual scenes by having the subject complete the last image in a series. The Tower of London Test and its variations involve a test subject moving colored beads among different pegs on a pegboard in order to replicate the design of the examiner. Subjects are scored by the number of moves undertaken to replicate the examiner’s design with the fewest moves resulting in the best score. In both tests subjects learn the strategy being employed, which is then changed, and the new strategy needs to be discovered and applied. While working memory deficits in visual spatial tasks have been espe­cially noted in Parkinson disease, verbal memory tests using “working memory” have also proven problematic in Parkinson disease. There is no evidence for language deficits (i.e., aphasia) but rather apparent deficits in the executive function of planning speech. As expected and similar to previously described visual spatial testing, deficits are most notable with greater complexity in the task. Testing that requires multiple simultaneous levels of processing results in the greatest deficits. Patients appear to be less able than control subjects to develop strategies that integrate and pri­oritize information to perform more complex memory tasks. For example, when interference is interjected into a memory testing paradigm deficits become more clearly evident (Taylor, Saint-Cyr, and Lang 1990; Cooper, Sagar, and Sullivan 1993). There would appear to be an inability to main­tain information in “working memory” when extraneous information is being introduced.

It is evident that executive cognitive dysfunction occurs early in Par­kinson disease with the greatest involvement mediated through “working memory” in prefrontal areas. These executive deficits are more evident when tasks become more complicated and especially when new rule changes must be incorporated. The next section discusses the dopaminer­gic pathways believed to mediate these processes.