Arun L.W. Bokde

Classifying brain states and determining the discriminating activation patterns: Support Vector Machine on functional MRI data

In the present study, we applied the Support Vector Machine (SVM) algorithm to perform multivariate classification of brain states from whole functional magnetic resonance imaging (fMRI) volumes without prior selection of spatial features. In addition, we did a comparative analysis between the SVM and the Fisher Linear Discriminant (FLD) classifier. We applied the methods to two multisubject attention experiments: a face matching and a location matching task. We demonstrate that SVM outperforms FLD in classification performance as well as in robustness of the spatial maps obtained (i.e. discriminating volumes). In addition, the SVM discrimination maps had greater overlap with the general linear model (GLM) analysis compared to the FLD. The analysis presents two phases: during the training, the classifier algorithm finds the set of regions by which the two brain states can be best distinguished from each other. In the next phase, the test phase, given an fMRI volume from a new subject, the classifier predicts the subjects instantaneous brain state.

Biomarkers for Alzheimer's disease: academic, industry and regulatory perspectives

Advances in therapeutic strategies for Alzheimers disease that lead to even small delays in onset and progression of the condition would significantly reduce the global burden of the disease. To effectively test compounds for Alzheimers disease and bring therapy to individuals as early as possible there is an urgent need for collaboration between academic institutions, industry and regulatory organizations for the establishment of standards and networks for the identification and qualification of biological marker candidates. Biomarkers are needed to monitor drug safety, to identify individuals who are most likely to respond to specific treatments, to stratify presymptomatic patients and to quantify the benefits of treatments. Biomarkers that achieve these characteristics should enable objective business decisions in portfolio management and facilitate regulatory approval of new therapies.

Core candidate neurochemical and imaging biomarkers of Alzheimer's disease*

In the earliest clinical stages of Alzheimers disease (AD) when symptoms are mild, clinical diagnosis can be difficult. AD pathology most likely precedes symptoms. Biomarkers can serve as early diagnostic indicators or as markers of preclinical pathologic change. Candidate biomarkers derived from structural and functional neuroimaging and those measured in cerebrospinal fluid (CSF) and plasma show the greatest promise. Unbiased exploratory approaches, eg, proteomics or cortical thickness analysis, could yield novel biomarkers. The objective of this article was to review recent progress in selected imaging and neurochemical biomarkers for early diagnosis, classification, progression, and prediction of AD.

Cognitive and clinical characteristics of patients with amyotrophic lateral sclerosis carrying a C9orf72 repeat expansion: a population-based cohort study

Summary Background Amyotrophic lateral sclerosis (ALS) is a progressive neurodegenerative disease of upper and lower motor neurons, associated with frontotemporal dementia (FTD) in about 14% of incident cases. We assessed the frequency of the recently identified C9orf72 repeat expansion in familial and apparently sporadic cases of ALS and characterised the cognitive and clinical phenotype of patients with this expansion. Methods A population-based register of patients with ALS has been in operation in Ireland since 1995, and an associated DNA bank has been in place since 1999. 435 representative DNA samples from the bank were screened using repeat-primed PCR for the presence of a GGGGCC repeat expansion in C9orf72. We assessed clinical, cognitive, behavioural, MRI, and survival data from 191 (44%) of these patients, who comprised a population-based incident group and had previously participated in a longitudinal study of cognitive and behavioural changes in ALS. Findings Samples from the DNA bank included 49 cases of known familial ALS and 386 apparently sporadic cases. Of these samples, 20 (41%) cases of familial ALS and 19 (5%) cases of apparently sporadic ALS had the C9orf72 repeat expansion. Of the 191 patients for whom phenotype data were available, 21 (11%) had the repeat expansion. Age at disease onset was lower in patients with the repeat expansion (mean 56·3 [SD 8·3] years) than in those without (61·3 [10·6] years; p=0·043). A family history of ALS or FTD was present in 18 (86%) of those with the repeat expansion. Patients with the repeat expansion had significantly more co-morbid FTD than patients without the repeat (50% vs 12%), and a distinct pattern of non-motor cortex changes on high-resolution 3 T magnetic resonance structural neuroimaging. Age-matched univariate analysis showed shorter survival (20 months vs 26 months) in patients with the repeat expansion. Multivariable analysis showed an increased hazard rate of 1·9 (95% 1·1–3·7; p=0·035) in those patients with the repeat expansion compared with patients without the expansion Interpretation Patients with ALS and the C9orf72 repeat expansion seem to present a recognisable phenotype characterised by earlier disease onset, the presence of cognitive and behavioural impairment, specific neuroimaging changes, a family history of neurodegeneration with autosomal dominant inheritance, and reduced survival. Recognition of patients with ALS who carry an expanded repeat is likely to be important in the context of appropriate disease management, stratification in clinical trials, and in recognition of other related phenotypes in family members. Funding Health Seventh Framework Programme, Health Research Board, Research Motor Neuron, Irish Motor Neuron Disease Association, The Motor Neurone Disease Association of Great Britain and Northern Ireland, ALS Association.

Prediction of conversion from mild cognitive impairment to Alzheimer's disease dementia based upon biomarkers and neuropsychological test performance

The current study tested the accuracy of primary MRI and cerebrospinal fluid (CSF) biomarker candidates and neuropsychological tests for predicting the conversion from mild cognitive impairment (MCI) to Alzheimers disease (AD) dementia. In a cross-validation paradigm, predictor models were estimated in the training set of AD (N = 81) and elderly control subjects (N = 101). A combination of CSF t-tau/Aβ(1-4) ratio and MRI biomarkers or neuropsychological tests (free recall and trail making test B (TMT-B)) showed the best statistical fit in the AD vs. HC comparison, reaching a classification accuracy of up to 64% when applied to the prediction of MCI conversion (3.3-year observation interval, mean = 2.3 years). However, several single-predictor models showed a predictive accuracy of MCI conversion comparable to that of any multipredictor model. The best single predictors were right entorhinal cortex (prediction accuracy = 68.5% (95% CI (59.5, 77.4))) and TMT-B test (prediction accuracy 64.6% (95% CI (55.5, 73.4%))). In conclusion, short-term conversion to AD is predicted by single marker models to a comparable degree as by multimarker models in amnestic MCI subjects.

Functional interactions of the inferior frontal cortex during the processing of words and word-like stimuli.

The hypothesis that ventral/anterior left inferior frontal gyrus (LIFG) subserves semantic processing and dorsal/posterior LIFG subserves phonological processing was tested by determining the pattern of functional connectivity of these regions with regions in left occipital and temporal cortex during the processing of words and word-like stimuli. In accordance with the hypothesis, we found strong functional connectivity between activity in ventral LIFG and activity in occipital and temporal cortex only for words, and strong functional connectivity between activity in dorsal LIFG and activity in occipital and temporal cortex for words, pseudowords, and letter strings, but not for false font strings. These results demonstrate a task-dependent functional fractionation of the LIFG in terms of its functional links with posterior brain areas.

Multivariate deformation-based analysis of brain atrophy to predict Alzheimer's disease in mild cognitive impairment.

Automated deformation-based analysis of MRI scans can be used to detect specific pattern of brain atrophy in Alzheimers disease (AD), but it still lacks an established model to derive the individual risk of AD in at-risk subjects, such as patients with mild cognitive impairment (MCI). We applied principal component analysis to deformation maps derived from MRI scans of 32 AD patients, 18 elderly healthy controls and 24 MCI patients. Principal component scores were used to discriminate between AD patients and controls and between MCI converters and MCI non-converters. We found a significant regional pattern of atrophy (p<0.001) in medial temporal lobes, neocortical association areas, thalamus and basal ganglia and corresponding widening of cerebrospinal fluid (CSF) spaces (p<0.001) in AD patients compared to controls. Accuracy was 81% for CSF- and 83% for brain-based deformation maps to separate AD patients from controls. Nine out of 24 MCI patients converted to AD during clinical follow-up. Discrimination between MCI converters and non-converters reached 80% accuracy based on CSF maps and 73% accuracy based on brain maps. In a logistic regression model, principal component scores based on CSF maps predicted clinical outcome in MCI patients even after controlling for age, gender, MMSE score and time of follow-up. Our findings indicate that multivariate network analysis of deformation maps detects typical features of AD pathology and provides a powerful tool to predict conversion into AD in non-demented at risk patients.

Diagnostic power of default mode network resting state fMRI in the detection of Alzheimer's disease

Functional magnetic resonance imaging (fMRI) of default mode network (DMN) brain activity during resting is recently gaining attention as a potential noninvasive biomarker to diagnose incipient Alzheimers disease. The aim of this study was to determine which method of data processing provides highest diagnostic power and to define metrics to further optimize the diagnostic value. fMRI was acquired in 21 healthy subjects, 17 subjects with mild cognitive impairment and 15 patients with Alzheimers disease (AD) and data evaluated both with volumes of interest (VOI)-based signal time course evaluations and independent component analyses (ICA). The first approach determines the amount of DMN region interconnectivity (as expressed with correlation coefficients); the second method determines the magnitude of DMN coactivation. Apolipoprotein E (ApoE) genotyping was available in 41 of the subjects examined. Diagnostic power (expressed as accuracy) of data of a single DMN region in independent component analyses was 64%, that of a single correlation of time courses between 2 DMN regions was 71%, respectively. With multivariate analyses combining both methods of analysis and data from various regions, accuracy could be increased to 97% (sensitivity 100%, specificity 95%). In nondemented subjects, no significant differences in activity within DMN could be detected comparing ApoE ε4 allele carriers and ApoE ε4 allele noncarriers. However, there were some indications that fMRI might yield useful information given a larger sample. Time course correlation analyses seem to outperform independent component analyses in the identification of patients with Alzheimers disease. However, multivariate analyses combining both methods of analysis by considering the activity of various parts of the DMN as well as the interconnectivity between these regions are required to achieve optimal and clinically acceptable diagnostic power.

Correlated gene expression supports synchronous activity in brain networks

Cooperating brain regions express similar genes When the brain is at rest, a number of distinct areas are functionally connected. They tend to be organized in networks. Richiardi et al. compared brain imaging and gene expression data to build computational models of these networks. These functional networks are underpinned by the correlated expression of a core set of 161 genes. In this set, genes coding for ion channels and other synaptic functions such as neurotransmitter release dominate. Science, this issue p. 1241 Gene expression is more similar than expected by chance in brain regions that are functionally connected. During rest, brain activity is synchronized between different regions widely distributed throughout the brain, forming functional networks. However, the molecular mechanisms supporting functional connectivity remain undefined. We show that functional brain networks defined with resting-state functional magnetic resonance imaging can be recapitulated by using measures of correlated gene expression in a post mortem brain tissue data set. The set of 136 genes we identify is significantly enriched for ion channels. Polymorphisms in this set of genes significantly affect resting-state functional connectivity in a large sample of healthy adolescents. Expression levels of these genes are also significantly associated with axonal connectivity in the mouse. The results provide convergent, multimodal evidence that resting-state functional networks correlate with the orchestrated activity of dozens of genes linked to ion channel activity and synaptic function.

Multivariate network analysis of fiber tract integrity in Alzheimer’s disease

Axonal and dendritic integrity is affected early in Alzheimers disease (AD). Studies using region of interest or voxel-based analysis of diffusion tensor imaging data found significant decline of fractional anisotropy, a marker of fiber tract integrity, in selected white matter areas. We applied a multivariate network analysis based on principal component analysis to fractional anisotropy maps derived from diffusion-weighted scans from 15 AD patients, and 14 elderly healthy controls. Fractional anisotropy maps were obtained from an EPI diffusion sequence using parallel imaging to reduce distortion artifacts. We used high-dimensional image warping to control for partial volume effects due to white matter atrophy in AD. We found a significant regional pattern of fiber changes (p < 0.01) indicating that the integrity of intracortical projecting fiber tracts (including corpus callosum, cingulum and fornix, and frontal, temporal and occipital lobe white matter areas) was reduced, whereas extracortical projecting fiber tracts, including the pyramidal and extrapyramidal systems and somatosensory projections, were relatively preserved in AD. Effects of a univariate analysis were almost entirely contained within the multivariate effect. Our findings illustrate the use of a multivariate approach to fractional anisotropy data that takes advantage of the highly organized structure of anisotropy maps, and is independent of multiple comparison correction and partial volume effects. In agreement with post-mortem evidence, our study demonstrates dissociation between intracortical and extracortical projecting fiber systems in AD in the living human brain.