NINDS has always considered Parkinson's disease as a pressing concern, and recognizes it as a global disorder rather than just a movement disorder. About a year ago, at the request of Congress, our group of scientists developed a five-year research agenda and professional judgement budget. The current working group on the cognitive and emotional aspects of Parkinson's disease is part of the follow up to that research agenda with the goal of obtaining advice and guidance on the identification of the unmet scientific needs in this scientific area. The working group participants focused on the following issues: circuitry, epidemiology, assessment, treatment strategies, potential clinical trials, and innovative research approaches for cognitive and emotional aspects of Parkinson's.
James Parkinson recognized "Melancholy" as an important aspect of Parkinson's disease when he described the disease that bears his name. In the past 200 years, scientists have not been able to make much progress in understanding the epidemiology, neurobiology, and treatment of Parkinson's depression. Diagnostic criteria for depression in Parkinson's really haven't even been established, and estimates of incidence vary from 4 to 75 percent of Parkinson's patients. Prevalence studies in research centers find depression in 40 to 50 percent of patients. However, community-based studies tend to show 10 percent and below.
It is difficult to understand Parkinson's disease with depression because many of the symptoms of Parkinson's disease alone, such as cognitive dysfunction, psychomotor retardation, flat affect, and anergia overlap those of depression. Symptoms of other psychological disorders that occur in Parkinson's, such as apathetic syndromes, affective lability, anxiety, and problems with sleep, concentration, and attention also overlap depression and need to be characterized. Moreover, the majority of currently existing scales for depression tends to include those same aspects in their inventory, and therefore these scales do not discriminate between the two disorders.
Most basic to the study of depression in Parkinson's, though, is that both disorders have subtypes, layering heterogeneity on heterogeneity. This realization prompted the working group participants of the January 25-26, 2001 NIH meeting almost universally to recommend subtyping Parkinson's and depression and studying the effects of treatment on each subtype. What worked for one subtype might not be effective on another. Depression respond differently in someone with, say, Parkinson's caused by a stroke compared to depression for a person with idiopathic Parkinson's.
Over all, though, studies have shown increased risk of depression and symptoms of depression among certain subgroups of Parkinson's patients: the akinetic rigid, those with onset before age 50, those with more or less hemisphere involvement. In the pallidotomy study, almost everyone who developed post-pallidotomy depression had a history of depression prior to the surgery.
In general, there are three general theories of depression and Parkinson's disease: 1) the psychosocial stress of having a chronic disease and disability may cause depression; 2) Parkinson's medications may cause depression; 3) or the underlying Parkinson's disease may cause depression. First, the stress theory: it does not appear that depression is simply a reaction to the chronic disabling nature of Parkinson's. One study showed that about one patient out of 20 presented with depression and anxiety before being diagnosed with Parkinson's. Chronic, disabling diseases that don't affect the central nervous system, such as diabetes and arthritis have lower rates of depression than does Parkinson's. The relationship between disability and depression is weak, and is correlated more with relative loss for a particular individual than with absolute amount of disability. The pallidotomy study showed that when a Parkinson's patient is treated for his/her depression, often the motor symptoms improve. Conversely, if the pallidotomy improves motor symptoms, so the patients have lost much of their rigidity and tremor and now can rise from a chair, walk, patients who are depressed will say that the procedure didn't help them and they don't feel any better.
As for the second theory of medication-induced depression, dopamine agonists have been linked to delirium, agitation, restlessness, and other psychiatric symptoms, but there is only weak evidence linking them with depression. There are perhaps as many as 70 percent of patients with severe on/off syndromes with depression, particularly during the off state. Although it is hard to identify the underlying depression, anxiety, and the on/off states, there is no evidence that the depression is caused by the medication.
It is difficult to assess the third theory: that depression is a part of the Parkinson's disease process. Biological markers for depression such as shortened REM latency, non-suppression of cortisol in the DST, decreased serotonin markers in the CSF, and decrease platelet imipramine binding are not sensitive, specific, or useful. The working group participants recognized the need to develop novel biomarkers. There is, though, the effect of the degeneration of the monoamine afferent system. The sub-cortical neurons that contribute dopamine, norepinephrine, and serotonin relate directly to depressive and cognitive symptoms. They degenerate at different rates in Parkinson's disease, which would make the symptoms of depression in Parkinson's patients heterogeneous. For example, dopamine neurons from the ventral tegmental area are involved in reward, and their degeneration may cause such symptoms such as loss of motivation and apathy.
Disturbances in the frontal cortex-basal ganglia-thalamic loop of Parkinson's patients reveal that patients with Parkinson's and depression have a greater decrease of metabolism in the striatum and orbital frontal cortex than patients with Parkinson's alone. This disturbance is similar in patients with major depression. Data obtained from positron emission tomography (PET) studies reveal that there seem to be two distinct processes in Parkinson's patients also exhibiting cognitive and emotional disturbances. Patients with cognitive impairment and depression tend to show more metabolic decrease in the superior temporal lobe. Patients with depression have more metabolic decrease in the prefrontal cortex.
Findings from deep brain stimulation therapies may also elucidate the neuro-circuitry of Parkinson's and depression. For example, in a 60-year-old woman with bilateral stimulators placed in the sub-pallidus nucleus, turning on the left lead in the substantia nigra produced an immediate depressive reaction. When it was turned off, the depression abated. A PET study after the stimulation was turned on showed patterns of activation from the globus pallidus going to the anterior thalamus, the orbital frontal cortex going to the amygdala, and then, to the surprise of scientists, the right parietal lobe. In the midst of the depressive symptoms, the patient had a sense of falling, falling down in a deep pit.
Depression is not the only emotional effect of Parkinson's. Sleep problems also have profound effects. People may leave their jobs primarily because of fatigue. Anxiety disorders are hard to discriminate from Parkinson's symptoms. Is it really a panic attack, or a patient going "off" and becoming anxious? Is a patient who doesn't want to eat in front of people suffering from social phobia, or just embarrassed by the dyskinesias? People go to the doctor and say that they are depressed, but often they experiencing emotional incontinence and find that they cry all the time.
There is also a need to better understand how much hormones such as thyroid hormone and testosterone contribute to Parkinson's and to depression. Some cases in the literature and reports from physicians' practice indicate that people with what are called 'the dwindles' do remarkably well with testosterone. Such people don't really look depressed, but they lose weight, are not eating, and have no energy. The incidence of hypogonadism is not clear, but probably it is more frequent than hypothyroidism, and clinics testing for it have found a surprising number of people in the very low range. It would probably be beneficial to routinely test patients with a certain profile. Testosterone therapy cannot be applied across the board and supplemental testosterone can aggravate a prostatic cancer. Further, the normal hormone range is not established, and a screening and a registry on Alzheimer and Parkinson's patients would be needed with large numbers of blood samples from many different patients.
It is still not clear how scientists should look at depression and related symptoms in Parkinson's patients. They could import ideas developed in geriatrics or Alzheimer's disease. Or, are these disorders in Parkinson's different enough that it would be useful to look at the psychopathology and behavioral symptomatology in Parkinson's as a brand new field, to characterize mood disorders and various affective disturbances?
Basal ganglia loops with the cortex appear to be important in psychosis, depression, or anxiety, although that is not the traditional understanding of their function. The standard concept had been that the basal ganglia collected information from many cortical areas, fed it into the striatum, which then ultimately influenced the motor cortex. In general, the basal ganglia, then, performed some sensory/motor transformation. New evidence has indicated that there are more active loops than just the sensory motor loops, and recently researchers have been able to study them by using rabies virus.
Rabies virus is a wonderful tracer for looking at connections in the nervous system. It infects only neurons, doesn't move through glia, moves across synapses exclusively in a retrograde direction, in a time-dependent manner, and doesn't cause cell lysis. Essentially investigators could inject a cortical area with virus, and in two days the virus moves to label all of the neurons that feed into that cortical area. By the third day, the virus would label all the second order neurons that project to that area, and with rabies, even a third order labeling was possible. This technique is allowing scientists to look at the brain's wiring diagram, and at the interconnection between neurons.
Injections of rabies virus in specific areas of the pre-motor, motor and supplementary motor cortical areas demonstrate that each of these areas receives input from a distinct region of output nuclei from the basal ganglia. About 15 percent or less of the basal ganglia is devoted to the motor cortex, and an estimated 50 percent or less is devoted to motor control. That means that at least half of basal ganglia output is directed towards non-motor areas in the prefrontal cortex, infra-temporal cortex, and orbital frontal cortex. Moreover, the research indicates that the basal ganglia has the capacity to influence all the cortical areas in the frontal lobe, anterior to the central sulcus and some interesting regions posterior to it. Any change in basal ganglia function will have widespread consequences in a wide variety of functions in cortical areas, and over many different domains.
It is then possible to see that alterations in the skeleton diagram for the motor cortex would produce the cardinal symptoms of Parkinson's disease. Similar circuits exist for the higher order visual areas in basal ganglia. Therefore, the complexities of selective degeneration and partial treatment of the circuit show that when L-dopa is being used to treat the motor cortex, it may cause drug toxicity and visual hallucinations if the visual circuits are still intact. The lore is that in nursing homes when there is a fire alarm, the first patients out are those with Parkinson's disease. That would fit with recent evidence that the pathway from the central nucleus to the amygdala to the primary motor cortex is intact in many Parkinson's patients.
It is important to note that the areas of the basal ganglia that degenerate in Parkinson's are immediately adjacent to regions of the basal ganglia interconnected with the amygdala, and adjacent to regions interconnected with the prefrontal cortex. As the areas of degeneration spread, there will be a wide spectrum of changes, based on anatomical substrates, neighbor-to-neighbor relations, and in particular, on the complexity of this set of systems feeding into the motor system.
Depression and its effects have been studied extensively in cardiovascular and cerebral vascular disease. Its incidence is about the same as in Parkinson's disease-7 to 8 percent for full-blown depression and 40 to 50 percent of patients with depressive symptoms. It seems useful to apply some of the lessons learned to Parkinson's disease.
The first important issue is the definition of depression in the context of cardiovascular, cerebral or Parkinson's disease. It is difficult to untangle depression from other behavioral symptoms. Second, the definition depends on what investigators want to study. For example, the syndromic definitions that are currently in vogue in psychiatry do not have any bearing to outcome as it relates to cardiovascular disease. When researchers wanted to study outcome, the recurrence of heart attacks at 3 months, 6 months and 18 months, they found that the death rate was three or four times higher with just a modest number of depressive symptoms. Meeting the criteria for full-blown depression didn't increase the risk very much. The data on cardiac disease is clear cut and will lead to a large scale, 2,000-patient clinical trial indicating whether cognitive therapy, for example, will reduce mortality and cardiac events. Other data sets seem to indicate that depression, even 20 to 25 years prior to cardiac disease is a good predictor of who will get it. There has been no clinical trial that is testing whether treating depression can be a factor in reducing cardiac disease.
In applying these lessons to Parkinson's disease, it is important to identify what symptom factors affect the Parkinson's outcome, then target treatment for that group of symptoms to see which medications work. It is necessary to remember that scientists don't really know how drugs affect depression in Parkinson's. Anecdotal reports indicate that serotonin reuptake inhibitors (SSRIs) worsen it, but the research is retrospective. The fundamental assumption is that if a drug works in primary depression, it will work in the context of a medical illness, and that probably is incorrect. There is a need to reexamine the use of drugs such as selegiline and monoamine oxidase (MAO) inhibitors, which have been used for both depression and Parkinson's disease.
Treating Parkinson's disease may help us understand some of these neural circuits. A term being used now for looking at the neuroanatomy and neurocircuitry of depression is "frontal striatal dysfunction," which describes Parkinson's disease as much as it describes depression. One study showed that with pallidotomy (PDT), there is a change in hypochondriasis, hypomania score and with thalamotomy there is a change in depression and social introversion. The numbers of subjects in this study were very small, and patients were not randomized, but the study provides some clues of overlapping circuits in Parkinson's disease. There are a few imaging studies, mostly PET, showing that in Parkinson's disease, blood flow changes are in the medial frontal cortex, cingulate cortex and in the caudate and orbital interior regions, when compared to non-depressed patients and control subjects. These same areas are being implicated in primary depression, and several studies are beginning to demonstrate that the medial orbital frontal region is critical in depression. Moreover, the number of glial cells and the size of the neurons are smaller in depressed individuals in this brain region.
One critical area of function affected in that part of the brain is the ability to deal with a negative input. If a non-depressed person is performing a task, and the feedback is that she is not doing too well, most individuals subsequently improve their performance. In persons with depression, that single feedback seems to impair their ability to do well. They actually do worse. And that persists even after recovery from depression, so that it appears more like a trait and less like a state. The question is, how closely is it related to the interior frontal region dysfunction, and how much is it related to Parkinson's disease?
The cardiovascular health study demonstrates that depression was related to the occurrence and severity of silent strokes. A recent study with a large data set showed that the lesions present in the medial frontal cortex were related to depression. Perhaps the pathology may tie it all together.
Parkinson's disease itself is a heterogeneous disorder, and the depression associated with it is heterogeneous with many conditions contributing to depressive symptoms in Parkinson's. Some patients have adjustment disorders, which would be an inordinate psychological reaction to the development of the disease. Some patient become demoralized, since they become discouraged about the development of the disease, but it isn't full blown depression. There are many reported cases of Dysthymia in the literature as well as the "Dwindles". Some patients have bipolar disorder, and there may be a specific Parkinson's disease-induced mood disorder. In addition there are drug and mood changes, anxiety syndromes, and pathological fearfulness. In emotional incontinence, patients will cry or be excessively sentimental for no apparent reason and often think they are depressed, but really they are just crying and do not have full blown depression. Patient can appear depressed when they have dementia, apathetic states, or states of delirium which can occur with medications.
Depression sometimes predates the diagnosis of Parkinson's disease. It is possible that some people simply have depression prior to the onset of Parkinson's. Alternatively, depression may be associated with the disease. The Danish literature shows about a 4 to 8 year prodrome in which 40 percent of the patients were having psychological symptoms. One of the problems is that in addition to all the overlapping heterogeneity, there is overlap with every cognitive, vegetative and somatic feature of Parkinson's and depression, but in particular with the motor symptoms. In addition, psychiatric conditions such as apathy, anxiety, and emotional incontinence occur comorbidly with depression.
The Udall Center at John Hopkins has provided as its main function, a recruitment resource for a brain donation program. When patients enroll in the program, they will their brain to research upon their death, and during their life they receive a psychiatric, motor and cognitive evaluation every two years. The population so far has some 38 percent with no current symptomatic psychiatric disorder, 24 percent with depression, and 17 percent with anxiety disorders. Some have psychosis only, and some have psychosis with depression and anxiety. Looking at only depression, anxiety, and no current psychiatric syndrome, there are more women than men with depression, and in the "off" state, depressed people had more motor impairment.
When looking at the overlap, though, it becomes complex. Some of those with no current symptoms had had prior depressive episodes since the onset of Parkinson's and were in full remission. 0thers had recurrent depression but did well on antidepressants. Of those with depression, nine had recurrent major depression, five were in the midst of a single episode, and five others had sort of mood disorders with depressive features that didn't quite meet the full criteria for a DMS-IV major depression. Among those with anxiety disorders, there was a range from generalized anxiety disorder, specific phobia, panic disorder, social phobia, and people who didn't fit into any particular category but had prominent anxiety. Six of those with anxiety had major depressive syndromes in full or partial remission.
Patients were ranked using a scaled neuropsychiatric inventory. It taps 12 different neuropsychiatric phenomena--delusions, hallucinations, agitation, depression, anxiety, elation, apathy, disinhibition, irritability, motor, sleep, and appetite. This scale is good in assessing whether or not patients have affective disturbance, but it is not helpful for distinguishing between patients with anxiety or depressive disorders. There is an obvious need for better study of the psychiatric phenomena and their overlap in Parkinson's, and it may be useful to question whether the traditional DSM nosology is optimal for studies of psychopathology in Parkinson's disease. Perhaps scientists should use a symptom clusters system as a way of defining the psychopathology.
One of the basic questions about Parkinson's disease is whether depression presages dementia. In comparative studies of the cognitive profiles of dementia, researchers followed both Alzheimer's and Parkinson's patients who were not demented at baseline, then if they became demented, looked back at the earliest clinical features of dementia. The neurophysiological battery they used was a brief screen that is designed to diagnose dementia. After two years, 40 of the 450 Alzheimer's patients had become demented. They found that, based on tests of memory, naming and abstract reasoning, they could discriminate, at least on a group level, those who would and who would not become demented. The results have been independently replicated.
In a similar study, they were able to look prospectively at about 200 people with Parkinson's who were not demented at baseline. Looking back two years, patients' scores on the Hamilton depression scale did differentiate those who would have incident dementia. That study has also been replicated, and it is apparent that pre-existing depression is a leading edge of the dementia syndrome. The cognitive profile two years earlier was also much different for those who did and did not become demented. With a larger set of data, a repeat of the analysis showed that a representation of memory using the selective reminding test predicted dementia even better. That was an indication that early memory changes occur in Parkinson's patients who will become demented. It is possible that what differentiates these patients is the involvement of the basal ganglia-frontal loops mentioned earlier.
In a study comparing matched demented Alzheimer's and Parkinson's patients, investigators found that performance in the naming and delayed recall tests declined more rapidly for the Parkinson's patients, indicating that this represented a later change, closer to the onset of dementia, than it did for the Alzheimer's patients. Also, the Parkinson's patients performed worse all along on category fluency, which suggests a pre-existing cognitive deficit upon which the dementia was overlaid. The Alzheimer's group performed more poorly on delayed recognition, which is consistent with the idea of an encoding deficit in Alzheimer's that exists before the dementia is recognized. Neuropsychological tests are proving useful in identifying the differences in the profiles of Parkinson's disease and Alzheimer's dementia. This is however a very complex undertaking. There are cognitive changes in early Parkinson's disease, but they may or may not be related to the later dementia. Dementia may be overlaid on the cognitive changes of Parkinson's disease, or it might be a whole different process. Scientists need to be careful about characterizing every aspect of that cognitive change-depression-dementia spectrum.
Currently, it is fairly easy to discriminate Alzheimer's and Parkinson's clinically. However, this distinction becomes difficult with patients who have an ex-parametrical kind of presentation. Perhaps it is early psychosis, Alzheimer's, dementia with Lewy bodies, cortical basal ganglia degeneration, or frontal temporal dementia. Therefore, researchers cannot consider dementia in Parkinson's disease as a unitary entity. Many factors may contribute to the manifestation of dementia: it may be an exacerbation of the dopaminergic deficit, or dementia with Lewy bodies, or Alzheimer's overlaid on these conditions. Until investigators can look at neuropsychological batteries and do cluster analyses and subtype these dementias, a lot of these issues will remain unclear.
Comparisons of the effects of Parkinson's disease and Alzheimer's disease can supply information on brain circuitry. Some of the current work focuses on understanding the role of the basal ganglia (especially the striatum) and the medial temporal lobes (especially the hippocampus) in human associative learning. Damage in the striatal regions is typical of Parkinson's disease and damage in the hippocampal region is seen in the very early stages of Alzheimer's.
By using computational neural-network modeling, behavioral analyses of animals with experimentally induced brain lesions, and neuropsychological studies of memory-impaired clinical populations, investigators have found converging evidence that the hippocampus acts as an essential gateway to memory. It modifies representations of stimuli in order to reflect significant patterns between different stimuli and between stimulus and outcome.
In contrast, investigators have come to view the basal ganglia as a system for mapping these new representations to outcomes and behavioral responses. Since generalizations depend critically on how one has formed these representations, investigators expected to find that Alzheimer's patients would learn a task at normal speed, but their representations would not be normal, and they would be poor in generalizing what they had learned to any new situation. By contrast, Parkinson's patients would learn more slowly, but their learning would be normal, and they would be able to generalize what they had learned to other situations normally.
One clinical trial compared non-demented elderly subjects who had the very mildest degree of hippocampal and entorhinal atrophy with early stage Parkinson's patients using a task called "acquired equivalence." In short, the first phase of this test involved learning that two girl's faces out of a group of faces always prefer the same sort of pet fish. The second phase of the task involved one of the two girls preferring a particular type of pet dog. Because there is equivalence in mapping preferences, the idea is to transfer the equivalence and indicate that the second girl also prefers that type of dog.
The results on the initial task showed that the control subjects and the hippocampal atrophied subjects performed very similarly. As expected, the Parkinson's patients were much slower in learning the initial task. With phase 2, the Parkinson's patients generalized as well as the controls did, and the hippocampal atrophied Alzheimer's patients were significantly impaired, could not generalize at all, and had to start their learning from scratch.
This, and other work, indicated that Parkinson's patients, with their cortical striatal dysfunction, learn more slowly than controls, but do not show qualitative differences in what is learned. That contrasts with the early Alzheimer's patients with medial temporal lobe damage. These patients are unable to transfer familiar stimuli to novel recombinations. Studies using more complex patterns have confirmed that model and further studies in animals have shown similar patterns. Other data currently suggest that beyond the overall slowing, there seems to be a particular Parkinson's problem in switching attention from certain attributes to others. That finding shows some analogues to the motor switching problems seen in Parkinson's disease.
Statistical parametric maps of glucose metabolism of non-demented Parkinson's patients and age-matched normal controls have been used to look at differences in brain function. Surprisingly, many brain regions not typically involved in the classical motor loops, showed decrease in glucose metabolism, including decreases in the cortex, prefrontal areas, anterior cingulate, and posterior parietal regions. These changes have been validated through several imaging techniques. At present, researchers really have not identified the networks for cognition and affect, and they need more specific markers in order to conduct clinical trials to see if treatment makes a difference in cognition and affect symptoms.
In an effort to look at the network correlates of cognition in Parkinson's disease, researchers used a principal component technique called partial square brain behavior analysis. Neuropsychological measures such as the CVLT, the visual-spatial Hooper test, the Beck Depression Inventory were taken concurrent with PET. These measures correlated with voxels that constitute networks. They found that there was a significant covariance pattern with both the CVLT and the Hooper tests. The two are uncorrelated as tasks, but do have the common thread of measuring executive function. With a very relaxed threshold, there would be prefrontal abnormalities, but researchers were impressed that the parietal, posterior parietal, and parietal occipital abnormalities correlate in the network structure.
Another network came out as an independent principal component. It showed that the deactivation of the orbital frontal region, and probably the anterior cingulate to some degree, covaried with metabolic increases posteriorally. This network correlated beautifully with the Beck Depression Inventory, even though these patients were non-depressed but dysphoric.
These two sets of maps provide trait markers for two aspects of the Parkinsonian non-motor symptoms, one for memory or executive function, and one for affect. There is a sense that they are quite different statistically and can be used for assessing therapy.
In other studies, investigators have developed tasks that allowed the assessment of interventions. For instance, a motor task that involved moving from a central point to a point on a circle and back to the central point. The targets were simply on the circle in clockwise sequence. In the next phase, the learning task, the movement was the same, but the targets were out of order and the sequence had to be learned. The two could be subtracted, isolating the learning. By measuring with an intervention "on" and "off," researchers could compare the baseline motor function and learning in the two therapeutic states.
Comparison of the learning curve of early stage Parkinson's patients and normal controls showed a learning deficit for those with Parkinson's, although these patients showed no cognitive defect in the ordinary way. Learning was increased with pallidal deep brain stimulation, and trended downward with levodopa infusion. An SPN map shows no change with levodopa, and pallidal DBS shows changes in the cortex, bilaterally, dorsal lateral, prefrontal area, and posterior parietal. These areas are not particularly related to motor function, but are modified by the pallidal manipulation. In graphing brain-behavior relationship, DBS significantly moves it along the normal line from being at the low end toward normal range. Levodopa may make it a little bit worse, but does not achieve significance.
So in the end, to study cognition and affect in non-demented patients, investigators need baseline studies with metabolism, perfusion surrogates to understand the traits for these conditions. They ultimately will try to use neural network covariance mapping to understand normal cognitive processes, the changes occurring in Parkinson's patients, and changes attributed to intervention therapy.
This breakout group wanted to get a better feel for the real incidence and prevalence of Parkinson's disease and then focus on target subtypes in order to really explicate the physiology and anatomy of this disorder.
Incidence and prevalence of Parkinson's disease is uncertain, so a universal screen needs to be developed in order to first establish the overall incidence of extrapyramidal disorders that could be subdivided into idiopathic Parkinson's disease and other disorders. This approach would allow the identification of the true incidence of genetic Parkinson's disease and the risk factors influencing the disease such as toxic exposure, neuroleptic exposure, vascular causes and multi-infarct dementia.
Once this kind of data is obtained, investigators could look at the true incidence of concomitant changes, and perhaps see how many of these changes predated the motoric components. Such an approach would provide information on the incidence of depression. It might even be possible, with an extensive screen, to pull out preventive factors such as antioxidants.
There is a real need for additional information on the basic neuroanatomy, neurophysiology and neurobiology of Parkinson's disease. Clinical studies could benefit from the use of imaging techniques such as functional magnetic resonance (fMRI), PET, SPECT and related techniques. With new developments in MR and increased resolution, investigators could precisely map the anatomical circuit of the disorder. The working group participants recognized the need to support the development of concurrent EEG and fMRI recording techniques, and possibly MEG and fMRI. While fMRI provides good spatial resolution, there is a need to but for integration with the faster, but less spatially sophisticated, electrophysiological measures in order to advance computational modeling of neural activity.
Some studies on anatomy and circuitry would need to be done in animal models, particularly nonhuman primates, since their cortex is more similar to that of humans. Network data in humans would need to be compared to that in nonhuman primates.
Deep brain stimulation (DBS) needs further elucidation. The physiology of DBS in the globus pallidus and the subthalamic nuclei are not fully understood. Combination of DBS with neurobiological measurement might provide a useful approach to elucidate both pharmacological mechanisms and anatomical pathways of Parkinson's disease.
Investigators need to know more about the cortical basal ganglia loops, the frontal parietal circuits, the circuitry of working memory, and network dysfunction in Parkinson's patients. It would be useful to know more about the Parkinsonian brain at the onset of disease, in patients with cognitive disturbance, in patients with depression, and in patients who may have coexisting diseases such as Alzheimer's.
In the Alzheimer's research field, researchers have developed a brain template for describing the changes in the size of the cerebral cortex. The development of a similar template would be useful for Parkinson's disease. For patients who do not respond to pharmacological intervention for their frontal executive disturbance, it would be useful to know whether it is possible to activate the cells in the frontal cortex, or whether these cells are being irrevocably lost.
New techniques in MRI perfusion, with spin tagging in the neck, will produce precise measures of the global cerebral baseline state that is needed for Parkinson's patients. MRI diffusion tensor analysis, which measures water flow in white matter pathways, will allow investigators to actually identify and characterize particular neural paths and could be used to assess Parkinson's disease physiological connectivity.
Investigators need MRI-based cytoarchitectonic mapping in humans. For instance, in looking at one of the areas important for executive disturbances in Parkinson's, area 9, each individual patient's area 9 is slightly differently placed. They overlap, but with a technique using MRI to clarify the subtle architectonic areas, researchers could greatly reduce the variance in imaging and lesion data.
Medication, surgery, and treatment outcomes need to be correlated. Comprehensive assessments of Parkinson's patients need to measure how the patient actually functions in the world, including the cognitive changes, behavior/depression changes, and activities in daily living. Large data sets with multiple groups of patients are necessary in order to obtain significant findings. It is also necessary to include nested subsets of patients, such as sub thalamic nuclei versus globus pallidus stimulation patients, in studies of combined anatomical and cognitive/motoric function such as MRI studies and integrated MRI-electrophysiological techniques.
Network analysis is absolutely critical, because the brain does not function as islands of activation. There is a need to support computational modeling of neural circuits. A unique group of patients, those with very focal unilateral lesions in the basal ganglia resulting from hypertensive hemorrhages or infarctions of lenticulostriate arteries, provide an opportunity to dissect out some of the behavioral, anatomical, and physiological function of the basal ganglia cortical circuits. Network analysis also requires sophisticated mathematical analysis of large data sets and there is need to provide support for such activities. The approach of "independent component analysis" (ICA) needs further exploration for the analysis of large electrophysiological or MR imaging. Moreover, wavelet analysis approaches with the potential for extracting brief neural events should also be investigated.
Other Animal Models
Besides the primate models, transgenic mice offer the opportunity to carry out gene insertion studies in order to establish the genetic underpinnings of basal ganglia dysfunction. Moreover, transgenic mice could be used to study complex behaviors such as cognition, as well as study of the effects of aging on muscle physiology. Recently, a Drosophila Parkinson's disease model has been characterized and this model could be useful for genetic studies of the disease.
The working group participants focused primarily on the need for clinical trials that would help differentiate the depressive symptoms of Parkinson's disease alone as compared to Parkinson's disease co-morbid with depression. The pressing need is for a multi-center clinical trial to look at the wide range of Parkinson's patients I order to define the elements of Parkinson's disease and depression, and to address the treatment of depression. The trial would examine and classify subtypes of Parkinson's disease, and sub-syndromal symptoms of depression. For instance, where do people who have Hamilton scores between 18 and 10 fall in the disease process? Are they treatable? What effect on their disease course does it have if they are treated?
The trial would also look at the relationship of physical disability to depression in Parkinson's disease using a multi-center, double blind, placebo controlled design with an SSRI or newer medication. It would look at a large number of subjects, with spectra of depression and Parkinson's, including depressed mood, minimal depressive symptoms, and patients with cognitive deficits, provided that they could answer the scales. The trial would also use cognitive probes to look at neural networks, and use neuropsychological data. Outcome measures would be functional--quality of life, MRSF--to see if the patients were really getting better. Outcomes of specific cognitive tests would help researchers understand the neuropathology of Parkinson's disease. Anxiety scores would also be important.
The initial studies would help select sensitive and specific assessment tools for depression. For example, the Diagnostic Interview Schedule with a Hamilton attached (DISH) is easy to administer, takes about 25 minutes, gives a depression diagnosis, and is used by clinicians in other trials of medical illnesses. This approach might be useful for comparisons to depression in other conditions such as cardiovascular disease. With a clearer identification of the different subtypes of Parkinson's patients, researchers could use metabolic studies and PET to classify them. A systematic look at family histories of depression or Parkinson's disease and previous history of depression could allow the assessment of risk factors in the development of depression in Parkinson's patients. Long-term follow up would also be necessary to assess the natural course of depression.
The trial would focus mainly on pressing clinical questions. Do SSRIs make Parkinson's worse? Do they interact with Parkinson's medications such as eldepryl? Parkinson's disease is a condition that affects many neurotransmitters, and medications such as SSRIs to treat depression affect many different neurotransmitters, while other medications are almost entirely specific to particular neurotransmitters. Their effect in subtypes of Parkinson's may be different. Results of these clinical studies could be combined with imaging and animal studies to gain a real understanding of the neurochemical pathways involved in Parkinson's disease alone or co-morbid with depression.
The main outcome of the trial would be to develop simple, easy screening tools that could accurately define different subtypes. The secondary aim would be to assess the relationship of depression to disability, social support, and late versus early cognitive impairment.
It's important to look at psychosis in Parkinson's disease. There is a pressing need for treatment approaches that address the psychiatric aspects of Parkinson's disease. Deep brain stimulation and pallidotomy offer a unique opportunity for psychiatrists to look at the effects of different perturbations of the neural circuitry and their effects on emotions. Techniques that look at changes in activation and stimulation could identify neurocircuits, particularly when the effect can be turned off and on, potentially in a scanner.
Many patients are being treated with deep brain stimulation, and there is a need to develop measures of the non-motor aspects of Parkinson's disease such as anxiety and depression. It would be useful to understand the incidence of these non-motor impairments, and how they really affect therapeutic interventions. Patients who have rapid fluctuations in mood could be imaged when they are euphoric and depressed. By using levodopa to create different mood states and imaging them, researchers could understand the mechanisms of the underlying circuitry.
Animal models of both Parkinson's disease and depression would be useful and advance our understanding of both conditions. At present, there is no rodent model for Parkinson's disease, although there are rodent models of depression. The development of superimposed models of Parkinson's disease and depression would be useful. Moreover, there is also a need for a model of depression in the MPTP monkey.
Network modeling is also important. There is interaction of dopamine and serotonin in every cortical cell, and such mechanisms would tell a lot about depression and depression in dopamine-depleted patients. To what extent do people chronically treated with dopamine become sensitized at the cellular level? There are transporter knockout models that lack norepinephrine or serotonin, and researchers should investigate those effects on behavior in some of the MPTP models.
While there are numerous morphometry MR studies in depression, there are very few similar studies in patients with Parkinson's depression. Researchers now need to examine the orbital frontal cortex, assess different brain areas using diffusion MR and determine if there is a correlation with depression.
Investigators need a Parkinson's brain bank, with psychiatric and medical assessments. Parkinson's disease with depression is an effective model of how the brain can degenerate and cause depression. Various PET marker studies and some metabolic studies of Parkinson's and depression indicate that some similarities in disease progression as compared to major depression alone. Parkinson's disease differentially affects multiple neurotransmitter systems and this could make it a great model for studying the heterogeneity of depression in patients.
Finally, the working group participants recognized that depression and its subtypes are not well defined in Parkinson's disease. The most cost effective way to establish the prevalence of depression in Parkinson's disease would be to add questions to other large studies, like the NHANES study and assess its presence in the community.
Alexander I. TrÖster, Ph.D.
Dept. of Psychiatry & Beh.Sciences
University of Washington School of Medicine
Dept. of Psychiatry & Beh.Sciences
Constantine G. Lyketsos, MD, MHS
Department of Psychiatry
Department of Psychiatry
Johns Hopkins University
William McDonald, M.D.
Wesley Woods Health Center
Athina Markou, Ph.D.
The Scripps Research Institute
Patricia S. Goldman-Rakic, Ph.D.
Yale University School of Medicine
Laura Marsh, M.D.
Department of Psychiatry
Johns Hopkins Hospital
Mahlon R. DeLong, M.D.
Department of Neurology
Emory University School of Medicine
Robert T. Knight, M.D.
Department of Psychology
University of CA-Berkeley
Yaakov Stern, Ph.D.
John D. Gabrieli, Ph.D.
Department of Psychology
Peter L. Strick, Ph.D.
Center Neural Basis Cognition
University of Pittsburgh School of Medicine
Department of Neurobiology Psychiatry
Mark A. Gluck, Ph.D.
Center Molecular Behavioral Neuroscience
Alice M. Cronin-Golomb, Ph.D.
Department of Psychology
David Eidelberg, Ph.D.
North Shore University Hospital
Irene H. Richard, Ph.D.
Department of Neurology
University of Rochester
K.R. Krishnan, M.B., Ch.B.
Psychiatry and Behavioral Sciences
Ira R. Katz, Ph.D.
Department of Psychiatry
University of Pennsylvania
Emmeline Edwards, Ph.D.
Systems and Cognitive Neuroscience
National Institute of Neurological Disorders and Stroke (NINDS)
Gerald Fischbach, Ph.D.
National Institute of Neurological Disorders and Stroke (NINDS)
Judith A. Finkelstein, Ph.D.
Sensory/Motor Disorders of Aging
National Institute on Aging (NIA)
Steve Foote, Ph.D.
Division of Neuroscience and Basic Behavioral Sciences
National Institute of Mental Health
Cheryl Kitt, Ph.D.
Systems and Cognitive Neuroscience
National Institute of Neurological Disorders and Stroke (NINDS)Neuroscience Center, Room 2109
Eugene Oliver, Ph.D.
National Institute of Neurological Disorders and Stroke (NINDS)
Steve Zalcman, M.D.
Clinical Neuroscience Research Branch
Division of Neuroscience and Basic Behavioral Sciences
National Institute of Mental Health
Molly Wagster, Ph.D.
Neuropsychology of Aging
National Institute on Aging (NIA)
Last updated April 12, 2011