Cortical atrophy in AD patients (FDR corrected for multiple comparisons). 

Context in source publication

Context 1

... and night time behavior disorders, and appetite and eating disor ders. Questions were read to the caregiver exactly as written. If the caregiver did not understand the question, it was re peated with synonyms. After reading the question, the care giver was asked if he/she noticed the behavior of the patient. If the answer was negative, the next set of questions was asked. If the answer was positive, subquestions were read and possible answers were “yes” or “no” The caregiver was then instructed to state the frequency and severity of the be havior in that domain, regarding alterations revealed by the subquestions. Scores were calculated by multiplying the fre quency (1–4: rarely, sometimes, often, and very often) and the severity (1–3: mild, moderate, and severe). Based on a study by Aalten et al. 1 , we considered the following four subsyn dromes: apathy, hyperactivity (the sum of agitation, disin hibition, irritability, euphoria, and aberrant motor behavior scores), psychosis (delusions, hallucinations, and night-time behavior disturbances), and affective syndrome (depression and anxiety). CDR was based on a semi-structured interview and the classification was based on the guidelines. The data were obtained by a 3T MRI scanner (Philips Achieva, Best, Netherlands). A set of structural images was composed of the following sequences: a) sagittal high-resolution T1-weighted with gradient echo images that were acquired with TR/TE = 7/3.2 ms, field of view (FOV) = 240×240, and isotropic voxels of 1 mm 3 , b) coronal and axial Fluid-attenuated Inversion Recovery, T2-weighted imag es, anatomically aligned at the hippocampus with image pa rameters set to TR/TE/TI = 12000/140/2850, FOV = 220×206, voxels reconstructed to 0.45×0.45×4.00 mm 3 , and gap between slices set to 1 mm, c) coronal inversion recovery T1-weighted images with TR/TE/TI = 3550/15/400, FOV = 180×180, and voxels reconstructed to 0.42×0.42×3.00 mm 3 , and d) coronal multi-echo (5 echoes) T2-weighted images with TR/TE = 3300/30, FOV = 180×180, voxels reconstructed to 0.42×0.42×3.00 mm 3 . Cortical thickness was determined using FreeSurfer soft ware v.5.3. Measurements were performed according to the protocol suggested by Fischl and Dale 7 . First, the images were aligned to the Talairach and Tournoux atlas. Next, the im ages grayscale intensity was normalized and corrected for magnetic field inhomogeneity. The extra-cerebral voxels were removed by the skull-stripped protocol. Next, the voxels were labeled as GM, WM, or cerebral spinal fluid and two surfaces were created: pial and white 7,8 . Cortical thickness was calculat ed as the shortest distance between the pial and white surface at each vertex across the cortical mantle. For all analyses, a Gaussian filter with 10 mm full width at half maximum was used for smoothing the surface. To evaluate possible cortical thickness differences, we used a General Linear Model with age as regressor. In order to correct for multiple comparisons, a False Discovery Rate test (FDR) was applied over the t-score maps generated from group analysis, enabling the identifica tion of significant clusters. To assess possible correlations be tween NPS and cortical thickness of patients, we performed a Pearson correlation. All correlation analyses were corrected for multiple comparisons using FDR test. All procedures were in accordance with the ethi- cal standards of the responsible committee on human experimentation (institutional and national) and with the Helsinki Declaration of 1975, as revised in 2000. Informed consent was obtained from all patients. Demographic and cognitive results are shown in Table 1. The distributions of NPI scores for each neuropsychiatric syn drome are shown in Figure 1. We performed the analysis considering groups CDR 1 and CDR 2 as one group, since we hypothesized that both groups could represent a continuum regarding AD’s course: the more atrophy due to neurodegeneration, the more prominent are the neuropsychiatric symptoms. In fact, in our sample, CDR 2 group had much more neuropsychiatric symptoms (Total NPI score mean ± SD: 18.0 ± 13.49) than CDR 1 group (8.43 ± 4.37), p = 0.03. We found significant differences between cortical thick ness measures in AD patients compared with the normal elderly, as shown by cortical atrophy in the AD group in the temporal lobe areas, the bilateral entorhinal cortex, the left middle and left inferior temporal gyrus, the posterior regions of right superior temporal sulcus, the bilateral precuneus, the bilateral superior frontal gyrus, and the bilateral lateral orbitofrontal gyri. We also found significant differences in the bilateral insula, the isthmus of cingulate gyrus , the bilateral anterior cingulate cortex, the left posterior cingulate cortex, the bilateral supramarginal gyri, the right fusiform gyrus, the inferior parietal gyrus, and the superior parietal gyrus in both hemispheres. All results were FDR-corrected for multiple comparisons (Figure 2 and Tables 2 and 3). We found significant correlations between cortical thickness and affective syndrome (depression and anxiety) (Figure 3 and Table 4). All the other regressions (apathy, psy chosis, and hyperactivity) did not survive multiple compari son corrections. The significant differences found in the cortical thickness of the AD group and the normal elderly corresponds to the well-established cortical atrophy of AD patients, even in mild and moderate stages. According to the pattern of AD atrophy, earlier atrophic sites are related to temporal and pari etal lobes. As the degeneration progresses, areas of the fron tal lobe may be affected. One important issue is the overlapping of cognitive symp toms from depression or early dementia. This overlapping sometimes makes it difficult to distinguish affective syn drome and dementia. Depression may predict a faster cognitive deterioration; depressed patients with MCI have twice the risk of progressing to AD compared with non-de pressed patients. In this study, we found significant correlations between cortical structures in the right hemisphere and affective syn drome. Interestingly, some of the more correlated regions (in sula, lateral orbitofrontal, and temporal pole) are important parts of brain networks that process emotional information. Although insular functions are not fully understood, there is evidence that the insula, especially its anterior part, pro cesses emotional experiences. Based on the role of the in sula encoding interoceptive signals from the body’s internal milieu that reflect autonomic activity, Damasio argued that the insula is the location within the brain where subjective feelings of emotion are generated 9 . The anterior insula is also a key region of the SN, which is involved in detecting, inte grating, and filtering relevant interoceptive, autonomic, and emotional information. The SN, with the anterior insula as its integral outflow hub, assists target brain regions to generate behavioral responses to salient stimuli 10 . In a previous study of our group, we showed that connectivity in the right insula of AD patients was related to symptoms of agitation, disin hibition, irritability, euphoria, and aberrant motor behavior 4 . Moreover, task-based functional MRI studies showed that hy peractivity of anterior insula has been consistently implicated with anxiety disorders 11 . Takahashi et al. also found GM vol ume reductions in the left anterior insula in depression 12 . We also found a significant correlation between affec tive syndrome and the orbitofrontal cortex (a region involved with socioemotional processing) and the right temporal pole. Although temporal pole atrophy is more commonly seen in a temporal variant of frontotemporal dementia than in AD, dysfunctions in the right side of this region are associated with changes in personality and social behavior. Moreover, other affect-related problems in temporal pole atrophy include depression, irritability, apathy, and emotion al blunting. Increased associations between the left temporal lobe and depressive and anxiety symptoms are recurrent in the literature, despite the existence of controversies concern ing laterality. Mendez et al., for example, verified a correla tion between anxiety and dysthymia with hypoperfusion of the right temporal lobe 13 . Other studies have also examined the correlation between depression and a reduction in cortical thickness on the entorhinal cortex and the anterior cingu lated cortex 14 . Because of these findings, it is possible to note the relevance of correlations between affective syndrome and the frontal and temporal lobes. The most likely hypothesis to explain the pathophysiology of affective syndrome is based on the proposition that disturbances in the temporal and frontal lobes may influence brain network disruptions, which occurs, for instance, in the limbic system. These disruptions may disorganize serotonergic and dopaminergic neurotrans- mitters in frontal-subcortical pathways and impact behavior. Thus, depressive symptoms in MCI and AD patients become clinically relevant and reflect dysfunctions in the frontal lobe with associated behavioral manifestations – agitation, disin hibition, and restlessness 15 . Our findings of a correlation between affective syndrome and frontal atrophy are also supported by information from the literature. A post-mortem study demonstrated a reduction in the density of prefrontal cortical glial cells, as well as providing convincing evidence for diminished prefrontal neuronal size in patients with depression 16, 17 . A recent meta- analysis involving MRI studies in late-life depression reported significant volume reductions in the orbitofrontal cortex, putamen, and thalamus 18 . The authors hypothesized that dis ruption of networks linking these frontal-subcortical struc tures, as well as limbic networks, exert an important role in the pathophysiology of late-life depression. ...