Abstract

Parkinson disease (PD) is no longer considered a complex motor disorder characterized by Parkinsonism but rather a systemic disease with variegated non-motor deficits and neurological symptoms, including impaired olfaction, autonomic failure, cognitive impairment, and psychiatric symptoms. Many of these alterations appear before or in parallel with motor deficits and then worsen with disease progression. Although there is a close relation between motor symptoms and the presence of Lewy bodies (LBs) and neurites filled with abnormal -synuclein, other neurological alterations are independent of the amount of -synuclein inclusions in neurons and neurites, thereby indicating that different mechanisms probably converge in the degenerative process. Involvement of the cerebral cortex that may lead to altered behaviour and cognition are related to several convergent factors such as (a) abnormal -synuclein and other proteins at the synapses, rather than LBs and neurites, (b) impaired dopaminergic, noradrenergic, cholinergic and serotoninergic cortical innervation, and (c) altered neuronal function resulting from reduced energy production and increased energy demands. These alterations appear at early stages of the disease and may precede by years the appearance of cell loss and cortical atrophy.

1. Introduction

Parkinson disease (PD) is clinically characterized by a complex motor disorder known as parkinsonism and is manifested principally by resting tremor, slowness of initial movement, rigidity, and general postural instability. These symptoms are mainly due to the loss of dopaminergic neurons in the substantia nigra pars compacta, leading to reduced dopaminergic input to the striatum and accompanied by adaptive responses in the internal and external globus pallidus, subthalamus, thalamus and substantia nigra pars reticularis. Round, hyaline neuronal cytoplasmic inclusions called Lewy bodies (LBs) and enlarged aberrant neurites and threads are found in the Parkinsonian substantia nigra [1, 2]. In addition to the substantia nigra, other nuclei are involved such as the locus ceruleus, reticular nuclei of the brain stem, and dorsal motor nucleus of the vagus, as well as the basal nucleus of Meynert, the amygdala and the CA2 area of the hippocampus. LBs and aberrant neurites are also found in these locations [17]. Similar lesions but extended to the cerebral neocortex are characteristic of a closely-related disease named Dementia with Lewy bodies (DLB) [8]. PD and DLB are therefore considered Lewy body diseases (LBDs). Neuropathology and clinical aspects of DLB have been revised in detail elsewhere [9, 10].

LBs and neurites are composed of aggregates of normal, misfolded and truncated proteins, and ubiquitin, all of which are stored in the cytoplasm as nondegraded by-products of the degenerative process [1114]. The main component of LBs and aberrant neurites is -synuclein which is abnormally phosphorylated, nitrated and oxidized, has an abnormal crystallographic structure and abnormal solubility, and is prone to the formation of aggregates and insoluble fibrils [1523]. For this reason, LBDs are categorized as α-synucleinopathies.

Mutations (A53T, A30P, E46K) in the α-synuclein gene (SNCA; PARK1) are causative of autosomal dominant PD [2426]. In addition, triplication or duplication of the α-synuclein locus is associated with PD [2730]. Together, these observations lay bare the crucial role of α-synuclein in the pathogenesis of a percentage of familial cases of PD. Recent studies have shown that methylation of the human SNCA intron 1 reduces gene expression, and inhibition of DNA methylation activates SNCA expression. Methylation of SNCA intron 1 is reduced in substantia nigra, putamen and cerebral cortex in PD, suggesting activation of SNCA in PD [31]. α-synuclein also appears to be regulated posttranscriptionally as two microRNAs, mir-7 and mir-153, which bind specifically to the 3'-untranslated region of α-synuclein and downregulate its mRNA and protein levels [32]. The two microRNAs reduce endogenous expression of α-synuclein [32, 33]. Whether variation in the miRNA-433 binding site of fibroblast growth factor 20 confers risk for PD by overexpression of α-synuclein [34] requires further validation.

Mutations in other genes are also the origin of familial and, in some cases, sporadic PD. These include parkin (PARK2) [35], DJ1 (PARK7) coding for Parkinson disease protein 7 [36], PINK1 (PARK6) coding for PTEN-induced putative kinase 1 [37], LRRK2 (PARK8) coding for leucine-rich repeat kinase 2 [38, 39] and HTRA2 (PARK13) coding for HtrA serine peptidase 2: HtrA2 [40]. Another gene involved in familial PD is UCHL1 (PARK5) coding for ubiquitin carboxyl-terminal hydrolase L1 [41]. A strong association between galactocerebrosidase mutations and PD has recently been reported [42]. Additonal loci associated with autosomal or recessive PD have been described (see [4345]). An important point is that not all familial cases with PD due to parkin and LRRK2 mutations have LBs, although all of them have predominant degeneration of the substantia nigra pars compacta (see [46], for review). Therefore, PD cases due to parkin and LRRK2 mutations without LBs cannot be considered as instances of LBD. Yet mutations in PINK1 are associated with LB pathology similar to that seen in sporadic PD [47].

Systematic study of cases with LB pathology has prompted a staging classification of PD based on the putative progression of LB pathology from the medulla oblongata (and olfactory bulb) to the midbrain, diencephalic nuclei, and neocortex [4850]). Stage 1 is characterized by LBs and neurites in the dorsal IX/X motor nuclei and/or intermediate reticular zone; there is also myenteric plexus involvement. Stage 2 affects the medulla oblongata and pontine tegmentum and covers pathology of stage 1 plus lesions in the caudal raphe nuclei, gigantocellular reticular nucleus, and ceruleus-subceruleus complex; the olfactory bulb is also involved. Stage 3 refers to pathology of stage 2 plus midbrain lesions, particularly in the pars compacta of the substantia nigra. Stage 4 includes basal prosencephalon and mesocortex pathology (cortical involvement confined to the transentorhinal region and allocortex, and CA2 plexus) in addition to lesions in the midbrain, pons and medulla oblongata. Stage 5 extends to sensory association areas of the neocortex and prefrontal neocortex. Stage 6 includes, in addition, lesions in first order sensory association areas of the neocortex and premotor areas; occasionally there are also mild changes in primary sensory areas and the primary motor field.

Cases with Lewy pathology in the brain stem without clinical evidence of parkinsonism are considered premotor PD or incidental LBD [1, 2]. Risk factor profiles in incidental LBD and PD are similar thus further supporting the idea that iLBD represents preclinical PD, arrested PD or a partial syndrome [51].

Furthermore, several atypical cases have been reported, the majority of them not following a clear gradient of Lewy body pathology from the medulla oblongata to the neocortex. LBs in the amygdala are predominant or even unique in some cases, most of them related with Alzheimer’s disease and Down syndrome. For this reason, amygdala-predominant LB disease has been categorized as a distinct entity [52]. Finally, a few cases have been reported with predominant cortical LB pathology and discrete numbers of LBs in the brain stem [53, 54]. Atypical cases constitute from five to ten percent of total LBD victims [5558].

Cumulative clinical evidence reveals that olfactory dysfunction, dysautonomia, sleep fragmentation, rapid eye movement behaviour disorder, mood and anxiety disorders, and depression may precede parkinsonian symptoms in a number of patients with PD clinically characterized by parkinsonism [5970]. Whether these clinical symptoms are associated with LBs in selected regions of the central, autonomous and peripheral nervous systems is a matter of study.

The neuropathological substrates of selected nonmotor symptoms in PD have been examined in previous reports [7] and will be reviewed in the following paragraphs.

3. Loss of Olfaction and the Olfactory Bulb and Tract in PD

α-synuclein pathology affecting neurons and neurites occurs in the olfactory bulb and related olfactory nuclei at very early stages of cases with PD-related pathology [50, 7175]. It also occurs in cases of AD with amygdala-predominant LB pathology [76]. Double-labelling immunohistochemical studies have shown that dopamine- and somatostnin-positive cells are rarely affected, whereas mitral cells, calcium-binding protein- and substance-P-positive cells are vulnerable [77]. Olfactory bulb synucleinopathy has high specificity and sensitivity for Lewy body diseases, and it has been suggested that olfactory bulb biopsy might be considered to confirm diagnosis in PD [78], an indication not approved by others [79].

The number of intracytoplasmic and neuritic α-synuclein inclusions in the olfactory bulb and tract is low in the majority of cases, thus suggesting that α-synuclein aggregates as visualized in current histological preparations barely begin to explain the severity of olfactory decline. It may be hypothesized that as in other regions, olfactory alterations in PD are the result of more complicated settings resulting from several molecular deficits. Although no direct information is as yet available in PD, recent studies have yielded substantial data about the molecular pathology of the olfactory bulb. Despite the relatively high content of glucose-6-phosphate dehydrogenase (G6PD), NADPH-cytochrome P450 oxidoreductase, glutathione reductase (GR) and NADPH-diaphorase (NADPH-d) in the olfactory bulb of rodents [80], significant changes in carbonylation and nitration have been found in the olfactory bulbs of old mice [81]. Targets of oxidation in aged olfactory bulbs, as revealed by redox proteomics, are aldolase 1 and ferritin heavy chain [81]. The effects of aging on oxidative stress damage in the olfactory bulb have been further demonstrated in accelerated senescence-prone, short-lived (SAMP) mice when compared with accelerated senescence-resistant, longer lived (SAMR) strains [82]. Therefore, aging in the olfactory bulb is associated with increased oxidative stress and oxidative damage. Whether these modifications are augmented in PD is not known, but, needless to say, future work will help to increase our understanding of molecular alterations, other than those related to α-synuclein, in the olfactory bulb and tract in PD.

A different approach has brought about interesting results. A series of PD patients underwent [(11)C]methyl-4-piperidinyl propionate acetylcholinesterase brain PET emission tomography and olfactory testing with the University of Pennsylvania Smell Identification Test. The diagnosis of PD was confirmed by [(11)C]dihydrotetrabenazine vesicular monoamine transporter type 2 PET. Cholinergic denervation of the limbic archicortex was a more robust determinant of hyposmia than nigrostriatal dopaminergic denervation in subjects with moderately severe PD [83]. However, it is worth stressing that no apparent abnormalities in the cholinergic system appear to be present at stages 1 and 2 of Braak, and, therefore, cholinergic denervation of the limbic cortex is probably not the only factor accounting for olfactory disorder in early premotor stages of PD.

Disorders of olfaction also occur in familial PD but they appear to be more benign in certain familial cases, linked with LRRK2 or parkin mutations, than in sporadic PD [84, 85].

4. Dysautonomia and Autonomic Nervous System in PD

Early studies demonstrated the presence of LBs in the parasympathetic ganglia, sympathetic ganglia, and enteric nervous system in PD [86, 87]. LBs are consistently found in the hypothalamus, sympathetic (intermediodorsal nucleus of the thoracic cord and sympathetic ganglia) and parasympathetic system (dorsal vagal and sacral parasympathetic nuclei, and peripheral parasympathetic ganglia), and enteric plexus [8890]. Regarding central medullary autonomic areas, the number of catecholaminergic and serotoninergic neurons is not significantly reduced in PD, although raphe neurons decline in number with disease progression [91]. Neuropathological studies in large cohorts of neurologically unimpaired aged individuals have shown that the autonomic nuclei of the spinal cord and the peripheral autonomic nervous system are affected early on by LB pathology [74, 9294]. Finally, α-synuclein-immunoreactive inclusions are seen in neurons of the Meissner’s and Auerbach’s plexuses and in the corresponding axons projecting into the mucosa [56, 95].

Sensitive immunohistochemical methods to detect phosphorylated α-synuclein have revelealed multiorgan localization and gradient distribution of aberrant α-synuclein deposits. The highest densities occurred in the spinal cord, paraspinal sympathetic ganglia, vagus nerve, gastrointestinal tract and endocrine organs. Within the gastrointestinal tract, the lower esophagus and the submandibular gland had higher numbers of inclusions than the colon and rectum [96].

The cardiovascular autonomic system is also affected in PD, and alterations implicate both tyrosine hydroxylase-positive (extrinsic) and negative (intrinsic) nerves of the cardiac plexus [76, 97]. Functional studies have also demonstrated cardiac involvement in PD. [1231] meta-iodobenzylguanidine (MIBG) myocardial scintigraphy has shown reduced MIBG uptake in PD [98102]. Importantly, decreased MIBG uptake precedes neuronal loss in the sympathetic ganglia [103105]. Interestingly, accumulation of α-synuclein aggregates occurs in the distal axons of the cardiac sympathetic nervous system preceding that of neuronal somata or neurites in the paravertebral sympathetic ganglia, thus indicating a centripetal degeneration of the cardiac sympathetic nerve in PD [105]. Olfactory tests, polysomnographic studies and MIBG myocardial scintigraphy in combination may be used to discover early signatures of the disease [106]. Interestingly, cardiac sympathetic denervation precedes nigrostriatal loss in individuals bearing the E46K mutation in SNCA [107].

These observations point to an association between synuclein deposits and impaired function in the autonomic nervous system. But this does not imply a causal relationship between these events. Several aspects are still elusive and require further study. (i) Autonomic symptoms are not always present in PD. (ii) LB pathology in autonomic peripheral ganglia and plexus is not always associated with clinical symptoms. (iii) Little is known about the nature and composition of LBs in peripheral autonomic nervous system. (iv) No data are available about molecular changes preceding, or associated with, early and late stages of LB pathology in the autonomic peripheral nervous system. (v) Little is known about the alterations other than the accumulation of abnormal α-synuclein that might cause altered autonomic functions in DLBs.

5. Sleep Disorders

Sleep disorders including sleep fragmentation, REM sleep behaviour disorders, and complex paroxysmal nocturnal motor behavioral disorders are common in PD [108110], and they may precede by decades motor symptoms [111]. The neuropathological substrates are poorly understood although affected nuclei in the brain stem including the pedunculopontine nucleus probably play key roles [112]. The ventral visual stream appears involved in visuoperceptive alterations associated with REM disorders [113]. Finally, hypocretin (orexin) cell loss following an anterior to posterior gradient has been found in the hypothalamus of PD cases with disease progression [114, 115]. This is not clearly accompanied by constant decrease in the expression levels of CSF orexin [116]. Whether orexin correlates with sleep attacks and its action is mediated by dopamine receptors 3 needs validation [117].

6. Cognitive Impairment and the Cerebral Cortex in PD

Changes in personality and moderate or mild cognitive debilitation are found in PD. Neuropsychiatric alterations and cognitive decline may occur at early stages of parkinsonism suggesting that they are an integral part of PD from the beginning of the disease in some patients. Characteristically, symptoms are often subtle at the beginning and difficult to detect without neuropsychological tests, although they become aggravated with progression of the disease. Deficits mainly affect executive function including working memory and visuospatial capacity. These are often accompanied by anxiety and depression, and excessive daytime sleepiness probably related with sleep disturbances (see [118, 119], for review).

Certain studies have shown an association between cortical LBs and cognitive impairment [120123]. Yet other studies have not confirmed this assumption [57, 58, 124127]. Moreover, associated AD pathology has been suggested as an important cofactor in the progression of cognitive impairment in PD [58]. Additional studies have not clarified a predictive role of LBs in the occurrence of cognitive deficits [128, 129] although LB pathology correlates with visual hallucinations when present in the medial temporal lobe and visual areas [130, 131]. Statistical analysis reveals that α-synuclein aggregates in limbic regions are related to dementia in PD as well as to visual hallucinations when there is an underlying dementia [132].

Neuropathlogical studies in a large series have confirmed that staging of LB pathology is barely applicable to cognitive impairment and dementia. Only a percentage of cases showed a relationship between cortical LBs and cognitive impairment and dementia [127]. Taken together, these observations strongly indicate that cortical Lewy bodies are not per se causative of dementia, but rather indicators of aggregates of pathological synuclein. Other factors are probably more responsible of altered cognitive function in PD.

It must be stressed that tau phosphorylation and α-synuclein phosphorylation are increased in synaptic-enriched fractions of frontal cortex homogenates in PD in the absence of LBs in the same tissue samples [133]. This indicates early α-synuclein alterations at the synapse even in cases with no cognitive impairment [133]. Recent observations have further demonstrated the presence of small abnormal aggregates of α-synuclein at the synapses [134, 135]. Therefore, abnormal aggregates at the synapses in greater numbers than large cytoplasmic and neuritic aggregates (LBs and aberrant neurites, resp.) may account for impaired function in PD. These are important independent observations showing that synaptic pathology occurs in the absence of LBs and that the most common alteration in the cerebral cortex in PD is pathology at the synapses rather than the presence of LBs.

It is worth stressing that altered α-synuclein may result in altered protein-protein interactions leading to altered synaptic function. Although these modifications are barely understood as yet, it is worth stressing that abnormal interactions have been reported between α-synuclein and Rab3a, a protein involved in synaptic vesicle trafficking, Rab5, a protein involved in dopamine endocytosis, and Rab8, a protein engaged in transport [136]. Altered interactions have also been suggested between altered α-synuclein and phospholipase C (PLC 1), a signalling downstream step of metabotropic glutamate receptors [137].

In other words, LBs per se have no direct impact on clinical symptoms but other more subtle abnormalities are causative of impaired cortical function. In addition to abnormal accumulation of altered proteins at the cortical synapses, a series of convergent approaches may help to increase understanding of the different factors leading to impaired cerebral function in PD. Some of them relate to impaired dopaminergic, noradrenergic, cholinergic and serotoninergic innervation of the cerebral cortex; others, to intrinsic metabolic deficits.

Cognitive and executive deficits have been related, in part, to reduced dopaminergic innervation in the nigrostriatal and mesocortical dopaminergic systems compromising directly and indirectly, via alteration of the basal ganglia, prefrontal cortical function [138142]. [18F] FDOPA uptake is reduced in frontal association areas in later stages of PD [143]. However, altered cognitive performance is not clearly related with impaired dopaminergic innervations of the cerebral cortex at early stages of the disease. PET studies with [11C] NNC112 and [18F] FDOPA have not shown significant associations between D(1) receptor density in the frontal cortex and performance at early stages of PD, in spite of a significant association between reduced [18F] FDOPA uptake in the putamen and poor performance in cognitive tests [144]. Along the same lines, attenuated dopamine release has been observed in the dorsal caudate but not in the medial prefrontal cortex in early PD patients [145].

Yet nondopaminergic systems are known to be damaged in PD, including the monoaminergic cells of the locus ceruleus, serotoninergic neurons of the raphe and cholinergic neurons of the nucleus basalis of Meynert [146149].

Cholinergic deficits have been postulated as causative of frontal dysfunction in PD [150, 151]. Recent studies have shown early alteration of the cholinergic innervation of the cerebral cortex in PD which increases in cases with dementia, thus correlating impaired cholinergic innervation and cognitive impairment [152154].

In the same line, alteration of the serotonin transporter, as revealed by 123I-FP-CIT SPECT, has been observed in PD and with much more severe involvement in DLB, despite the comparable loss of striatal dopamine transporter [155].

Besides the loss of afferencies, primary impaired metabolism of the cerebral cortex may be causative of intrinsic cortical decay.

Cerebral glucose metabolism is reduced in the cerebral cortex in PD patients suffering from cognitive impairment [156]. Limited, mainly posterior, blood flow reductions have been reported in PD cases with mild cognitive deficits assessed by rCB scintigraphic study using TC-HMPAO-SPECT [157]. Metabolic and neuroimaging observations have recently documented decreased prefrontal and parietal 18F-fluorodeoxyglycose uptake in PD cases with mild cognitive deficits [158, 159]. Parallel conclusions have been obtained using magnetic resonance; T1-weighed images and mean diffusitivity and fractional anisotropy values are increased in the frontal cortex in PD [160, 161]. White matter hyperintensities are more frequent in PD cases with altered cognition than in cases with preserved cognitive functions [162]. Yet vascular abnormalities are very common in aged patients with PD [163], and we cannot rule out the possibility that white matter alterations are related to associate vascular/circulatory lesions rather than to primary lesions of PD.

A detailed discussion of molecular events leading to intrinsic cortical deficiencies is provided in the following paragraph.

7. Mitochondria and Energy Machinery Failure in the Cerebral Cortex in PD

Classical studies revealed abnormalities in complex I of the respiratory chain in the substantia nigra in sporadic PD (see [164], for review). More recently, several genes encoding proteins relevant to maintaining mitochondrial integrity have been shown to be causative of familial PD [165, 166]. These data further reinforce the role of mitochondria in the pathogenesis of PD. Interestingly, several mutant proteins associated with familial PD are linked to mitochondria [167]. DJ1 is localized in the mitochondria and modulates responses to oxidative stress [168, 169]. PINK1 is a protein kinase localized in the mitochondria in which mutations in the kinase domain are associated with mitochondrial deficits [170]. Furthermore, PINK1 is required for mitochondrial function as it interacts with and complements parkin [171173]. LRRK2 is a kinase localized in the outer mitochondrial membrane [174]. Finally, HtrA2 is localized in the mitochondria and is involved in apoptosis [175]. Deficits in mitochondrial function have also been identified in patients with DJ1, parkin and PINK1 mutations [176]. Interaction of these different gene products seems necessary to maintain mitochondrial homeostasis [177179]. Moreover, mutations in mitochondrial DNA have also been noted in familial parkinsonism due to PINK1 mutations [180].

These observations point to the possibility that mitochondrial dysfunction plays a crucial role not only in dopaminergic neurons of the substantia nigra but also in the whole brain. In this line, brain cortex and mitochondrial O2 uptake and complex I activity are significantly lower in PD, whereas mitochondrial nitric oxide synthase activity, cytochrome content, expression of Mn-superoxide dismutase (SOD2), mitochondrial mass, and oxidative damage are significantly higher in the frontal cortex in PD. The decreases in tissue and mitochondrial 02 uptake and in complex I activity are considered the consequences of mitochondrial oxidative damage in the cerebral cortex in PD [181, 182]. Moreover, subunits of mitochondrial complex I are oxidatively damaged, functionally impaired and misassembled in PD [183]. Finally, phosphorus and proton magnetic resonance spectroscopy confirm generalized mitochondrial dysfunction in PD [184]. It is not reckless to assume that loss of mitochondrial function is a primary cause of energy production decay.

Increased oxidative damage has also been detected in the frontal cortex, in addition to that reported several years ago in the substantia nigra, in PD [185]. Several key proteins are targets of oxidative damage in the frontal cortex even at very early stages of PD-related pathology, including α-synuclein, β-synuclein and SOD2 [186, 187]. Other relevant proteins are also oxidatively damaged in PD: UCHL1, Cu, Zn-superoxide dismutase and DJ-1 [188190]. In addition, increased oxidative damage of aldolase A, enolase 1 and glyceraldehyde dehydrogenase (GAPDH), all of them involved in glycolysis and energy metabolism, is found in the frontal cortex in premotor stages of PD and in established parkinsonian PD disease as well [191]. In the same line of generalized oxidative stress and stress responses is the observation of increased glutahione peroxidase, one of the main antioxidant enzymes inactivating hydrogen peroxide, in microglial cells of the gray matter and white matter in PD and DLB [192].

Recent observations have shown abnormal lipid composition in the frontal cortex at very early stages of PD-related pathology with significantly increased expression levels of the highly peroxidizable docosahexanoic acid (DHA) and increased peroxidability index [186]. Together, these features indicate that mitochondrial abnormalities, altered lipid composition and increased oxidative damage of proteins involved in the cytoskeleton, neurotransmission, mitochondrial function and energy metabolism occur at early, premotor stages of PD and persist with disease progression [193]. In the same line, recent observations have shown impaired lipid composition of lipid rafts with dramatic reductions in their contents of n-3 and n-6 LCPUFA, especially docosahexaenoic acid (22:6-n3) and arachidonic acid (20:4n-6), and increased saturated fatty acids (16:0 and 18:0) when compared with control brains, thus leading to increased membrane viscosity and, probably, to increased energy demands [194].

The term “exhausted neuron” was employed in Alzheimer’s disease to designate a combination of metabolic events leading to impaired and persistent energy production accompanied by increased energy demands that may be detected at very early stages of disease even in cases without overt clinical symptoms of cognitive impairment and dementia [195]. A similar scenario also occurs, albeit with different targets (different primary involvement of respiratory chain complexes, different lipoxidative and glycoxidative damage, different alteration of membrane lipid composition), in PD. It may be postulated that intrinsic exhaustion of neurons plays an important role in the subtle but inexorable progression of clinical symptoms once thresholds of neuronal tolerance cannot longer support energy demands.

Oxidative damage in cerebral cortex has also been reported in familial cases bearing the LRRK2 mutation in the absence of apparent cognitive impairment and in the absence of LBs in cerebral cortex [196]. This further supports the concept that the cerebral cortex is affected in PD independently of its being associated (or not) with the presence of LBs.

Altered metabolic events in the cerebral cortex in Parkinson disease are summarized in Table 1.

8. Psychiatric Symptoms and Neuropathological Correlates

Anhedonia, apathy, anxiety, panic attacks, social phobias and depression also occur in patients with PD even at early premotor stages and not related to medication [68, 197, 198]. Psychotic symptoms are frequent such as visual and auditory hallucinations, agitated confusion, vivid dreaming, delirium and delusions [199203]. The molecular substrates of such alterations are scarcely known but several hypothesis have been proposed including imbalance between serotoninergic and dopaminergic systems, cortical cholinergic deficiency and overstimulation of mesocorticolimbic dopamine receptors [201, 203207]. Neuropathological studies have helped little to increase understanding of depression and psychoses although recent observations have suggested that depression is related more to catecholaminergic than serotoninergic dysfunction [208], and that hallucinations correlate with the number of Lewy bodies in the temporal lobe, claustrum and visual cortex [130, 131].

Finally, an interesting and not fully understood paradigm is the consequence of alterations in amygdala in PD in spite of its constant involvement in classical PD and its almost unique alteration in amygdala-predominant LBD. The amygdala is activated in appetitive and emotional learning [209, 210]. Yet decreased responsiveness is found in the amygdala in PD in the face of fearful facial expressions, facial, prosodic and written verbal stimuli, and decision-making and facial emotion recognition [211213]. Whether these modifications are the result of impaired dopaminergic regulation [214] or the consequence of primary pathology in the amygdala is not known.

Acknowledgments

Work carried out in the Institute of Neuropathology was partially funded by Grants from the Spanish Ministry of Health, Instituto de Salud Carlos III PI05/1570, PI05/2214 and PI08/0582, PET2007-0397, and supported by the European Commission: BrainNet Europe II, LSHM-CT-2004-503039 and INDABIP FP6-2005-LIFESCIHEALTH-7 Molecular Diagnostics. Thanks are due to T. Yohannan for the editorial help.