A Machine Learning-Based Holistic Approach to Predict the Clinical Course of Patients within the Alzheimer's Disease Spectrum.

BACKGROUND: Alzheimer's disease (AD) is a neurodegenerative condition driven by multifactorial etiology. Mild cognitive impairment (MCI) is a transitional condition between healthy aging and dementia. No reliable biomarkers are available to predict the conversion from MCI to AD. OBJECTIVE: To evaluate the use of machine learning (ML) on a wealth of data offered by the Alzheimer's Disease Neuroimaging Initiative (ADNI) and Alzheimer's Disease Metabolomics Consortium (ADMC) database in the prediction of the MCI to AD conversion. METHODS: We implemented an ML-based Random Forest (RF) algorithm to predict conversion from MCI to AD. Data related to the study population (587 MCI subjects) were analyzed by RF as separate or combined features and assessed for classification power. Four classes of variables were considered: neuropsychological test scores, AD-related cerebrospinal fluid (CSF) biomarkers, peripheral biomarkers, and structural magnetic resonance imaging (MRI) variables. RESULTS: The ML-based algorithm exhibited 86% accuracy in predicting the AD conversion of MCI subjects. When assessing the features that helped the most, neuropsychological test scores, MRI data, and CSF biomarkers were the most relevant in the MCI to AD prediction. Peripheral parameters were effective when employed in association with neuropsychological test scores. Age and sex differences modulated the prediction accuracy. AD conversion was more effectively predicted in females and younger subjects. CONCLUSION: Our findings support the notion that AD-related neurodegenerative processes result from the concerted activity of multiple pathological mechanisms and factors that act inside and outside the brain and are dynamically affected by age and sex.

Involvement of ordinary what and where auditory cortical areas during illusory perception.

The focus of the present study is on the relationships between illusory and non-illusory auditory perception analyzed at a biological level. To this aim, we investigate neural mechanisms underlying the Deutsch's illusion, a condition in which both sound identity ("what") and origin ("where") are deceptively perceived. We recorded magnetoencephalogram from healthy subjects in three conditions: (a) listening to the acoustic sequence eliciting the illusion (ILL), (b) listening to a monaural acoustic sequence mimicking the illusory percept (MON), and (c) listening to an acoustic sequence similar to (a) but not eliciting the illusion (NIL). Results show that the areas involved in the illusion were the Heschl's gyrus, the insular cortex, the inferior frontal gyrus, and the medial-frontal gyrus bilaterally, together with the left inferior-parietal lobe. These areas belong to the two main auditory streams known as the what and where pathways. The neural responses there observed indicate that the sound sequence eliciting the illusion is associated to larger activity at early and middle latencies and to a dynamic lateralization pattern net in favor of the left hemisphere. The present findings extend to illusory perception the well-known what-where auditory processing mechanism, especially as regards tardy latency activity.

Thalamic Involvement in Fluctuating Cognition in Dementia with Lewy Bodies: Magnetic Resonance Evidences.

Dementia with Lewy bodies (DLB) is characterized by fluctuation in cognition and attention. Thalamocortical connectivity and integrity of thalami are central to attentional function. We hypothesize that DLB patients with marked and frequent fluctuating cognition (flCog) have a loss of thalamocortical connectivity, an intrinsic disruption to thalamic structure and imbalances in thalamic neurotransmitter levels. To test this, magnetic resonance imaging (MRI), diffusion tensor imaging (DTI) and proton MR spectroscopy on thalami were performed on 16 DLB, 16 Alzheimer's disease (AD) and 13 healthy subjects. MRI and DTI were combined to subdivide thalami according to their cortical connectivity and to investigate microstructural changes in connectivity-defined thalamic regions. Compared with controls, lower N-acetyl-aspartate/total creatine (NAA/tCr) and higher total choline/total creatine (tCho/tCr) values were observed within thalami of DLB patients. tCho/tCr increase was found within right thalamus of DLB patients as compared with AD. This increase correlated with severity and frequency of flCog. As compared with controls, DLB patients showed bilateral damage within thalamic regions projecting to prefrontal and parieto-occipital cortices, whereas AD patients showed bilateral alteration within thalamic region projecting to temporal cortex. We posit that microstructural thalamic damage and cholinergic imbalance may be central to the etiology of flCog in DLB.

The sound of consciousness: neural underpinnings of auditory perception.

The neural correlates of consciousness (NCC), i.e., patterns of brain activity that specifically accompany a particular conscious experience, have been investigated mainly in the visual system using particularly suited paradigms, such as binocular rivalry and multistable percepts in combination with neural recordings or neuroimaging. Through the same principles, we look here for possible NCC in the auditory modality exploiting the properties of the Deutsch's illusion, a stimulation condition in which a sequence of two specular dichotic stimuli presented in alternation causes an illusory segregation of pitch and side (ear of origin), which can yield up to four different auditory percepts per dichotic stimulus. Using magnetoencephalography in humans, we observed cortical activity specifically accompanying conscious experience of pitch inside an early bilateral network, including the Heschl's gyrus, the middle temporal gyrus, the right inferior, and the superior frontal gyri. The conscious experience of perceived side was instead accompanied by later activity observed bilaterally in the inferior parietal lobe and in the superior frontal gyrus. These results suggest that the NCC are not independent of stimulus features and modality and that, even at the higher cortical levels, the different aspects of a single perceptual scene may not be simultaneously processed.

Neuromagnetic functional coupling during dichotic listening of speech sounds.

The present magnetoencephalography (MEG) study tested the hypothesis of a phase synchronization (functional coupling) of cortical alpha rhythms (about 6-12 Hz) within a "speech" cortical neural network comprising bilateral primary auditory cortex and Wernicke's areas, during dichotic listening (DL) of consonant-vowel (CV) syllables. Dichotic stimulation was done with the CV-syllable pairs /da/-/ba/ (true DL, yielded by stimuli having high spectral overlap) and /da/-/ka/ (sham DL, obtained with stimuli having poor spectral overlap). Whole-head MEG activity (165 sensors) was recorded from 10 healthy right-handed non-musicians showing right ear advantage in a speech DL task. Functional coupling of alpha rhythms was defined as the spectral coherence at the following bands: alpha 1 (about 6-8 Hz), alpha 2 (about 8-10 Hz), and alpha 3 (about 10-12) with respect to the peak of individual alpha frequency. Results showed an inverse pattern of functional coupling: during DL of speech sounds, spectral coherence of the high-band alpha rhythms increased between left auditory and Wernicke's areas with respect to sham DL, whereas it decreased between left and right auditory areas. The increase of functional coupling within the left hemisphere would underlie the processing of the syllable presented to the right ear, which arrives to the left auditory cortex without the interference of the other syllable presented to the left ear. Conversely, the decrease of inter-hemispherical coupling of the high-band alpha might be due to the fact that the two auditory cortices do not receive the same information from the ears during DL. These results suggest that functional coupling of alpha rhythms can constitute a neural substrate for the lateralization of auditory stimuli during DL.