Assessing clinical utility of machine learning and artificial intelligence approaches to analyze speech recordings in multiple sclerosis: A pilot study.

BACKGROUND: An early diagnosis together with an accurate disease progression monitoring of multiple sclerosis is an important component of successful disease management. Prior studies have established that multiple sclerosis is correlated with speech discrepancies. Early research using objective acoustic measurements has discovered measurable dysarthria. METHOD: The objective was to determine the potential clinical utility of machine learning and deep learning/AI approaches for the aiding of diagnosis, biomarker extraction and progression monitoring of multiple sclerosis using speech recordings. A corpus of 65 MS-positive and 66 healthy individuals reading the same text aloud was used for targeted acoustic feature extraction utilizing automatic phoneme segmentation. A series of binary classification models was trained, tuned, and evaluated regarding their Accuracy and area-under-the-curve. RESULTS: The Random Forest model performed best, achieving an Accuracy of 0.82 on the validation dataset and an area-under-the-curve of 0.76 across 5 k-fold cycles on the training dataset. 5 out of 7 acoustic features were statistically significant. CONCLUSION: Machine learning and artificial intelligence in automatic analyses of voice recordings for aiding multiple sclerosis diagnosis and progression tracking seems promising. Further clinical validation of these methods and their mapping onto multiple sclerosis progression is needed, as well as a validating utility for English-speaking populations.

Ultra-violet light induced changes in DNA dynamics may enhance TT-dimer recognition.

Short-wave ultra-violet light promotes the formation of DNA dimers between adjacent thymine bases, and if unrepaired these dimers may induce skin cancer. Living cells have a very robust repair system capable of repairing hundreds of lesions every day. Although many of the details of the dimer repair mechanism are known, it is still a mystery how the dimers are recognized. Because the dimers are hidden from repair proteins diffusing in the cell nucleus, it has been surmised that dimer recognition is indirect. In this paper, a new recognition signal is suggested by a theory of the dimer-induced large amplitude, prolonged oscillations in the motion of the two strands in double-stranded DNA molecules. These large amplitude oscillations of the two DNA strands, localized around the dimer will unveil the dimer allowing the repair proteins to bind to the dimer site. The temperature dependence of the recognition rate is correlated with the inter-strand fluctuations and must decrease with decreasing temperature according to the findings in this paper. Moreover the probability for finding a large opening is localized to the dimer neighbourhood and these large openings may play an important role in dimer-repair protein biochemistry.

Bubble statistics and dynamics in double-stranded DNA.

The dynamical properties of double-stranded DNA are studied in the framework of the Peyrard-Bishop-Dauxois model using Langevin dynamics. Our simulations are analyzed in terms of two distribution functions describing localized separations ("bubbles") of the double strand. The result that the bubble distributions are more sharply peaked at the active sites than thermodynamically obtained distributions is ascribed to the fact that the bubble lifetimes affect the distributions. Certain base-pair sequences are found to promote long-lived bubbles, and we argue that this is a result of length scale competition between the nonlinearity and disorder present in the system.

Aging and immortality in a cell proliferation model.

We investigate a model of cell division in which the length of telomeres within a cell regulates its proliferative potential. At each division, telomeres undergo a systematic length decrease as well as a superimposed fluctuation due to exchange of telomere DNA between the two daughter cells. A cell becomes senescent when one or more of its telomeres become shorter than a critical length. We map this telomere dynamics onto a biased branching-diffusion process with an absorbing boundary condition whenever any telomere reaches the critical length. Using first-passage ideas, we find a phase transition between finite lifetime and immortality (infinite proliferation) of the cell population as a function of the influence of telomere shortening, fluctuations, and cell division.

Modelling the magnetic signature of neuronal tissue.

Neuronal communication in the brain involves electrochemical currents, which produce magnetic fields. Stimulus-evoked brain responses lead to changes in these fields and can be studied using magneto- and electro-encephalography (MEG/EEG). In this paper we model the spatiotemporal distribution of the magnetic field of a physiologically idealized but anatomically realistic neuron to assess the possibility of using magnetic resonance imaging (MRI) for directly mapping the neuronal currents in the human brain. Our results show that the magnetic field several centimeters from the centre of the neuron is well approximated by a dipole source, but the field close to the neuron is not, a finding particularly important for understanding the possible contrast mechanism underlying the use of MRI to detect and locate these currents. We discuss the importance of the spatiotemporal characteristics of the magnetic field in cortical tissue for evaluating and optimizing an experiment based on this mechanism and establish an upper bound for the expected MRI signal change due to stimulus-induced cortical response. Our simulations show that the expected change of the signal magnitude is 1.6% and its phase shift is 1 degrees . An unexpected finding of this work is that the cortical orientation with respect to the external magnetic field has little effect on the predicted MRI contrast. This encouraging result shows that magnetic resonance contrast directly based on the neuronal currents present in the cortex is theoretically a feasible imaging technique. MRI contrast generation based on neuronal currents depends on the dendritic architecture and we obtained high-resolution optical images of cortical tissue to discuss the spatial structure of the magnetic field in grey matter.

Pinning frequencies of the collective modes in alpha-uranium.

Uranium is the only known element that features a charge-density wave (CDW) and superconductivity. We report a comparison of the specific heat of single-crystal and polycrystalline alpha-uranium. In the single crystal we find excess contributions to the heat capacity at 41 K, 38 K, and 23 K, with a Debye temperature ThetaD = 265 K. In the polycrystalline sample the heat capacity curve is thermally broadened (ThetaD = 184 K), but no excess heat capacity was observed. The excess heat capacity Cphi (taken as the difference between the single-crystal and polycrystal heat capacities) is well described in terms of collective-mode excitations above their respective pinning frequencies. This attribution is represented by a modified Debye spectrum with two cutoff frequencies, a pinning frequency V0 for the pinned CDW (due to grain boundaries in the polycrystal), and a normal Debye acoustic frequency occurring in the single crystal.

Coexistence of superconductivity and ferromagnetism in ferromagnetic metals.

We address the question of coexistence of superconductivity and ferromagnetism. Using a field theoretical approach we study a one-fermion effective model of a ferromagnetic superconductor in which the quasiparticles responsible for the ferromagnetism form the Cooper pairs as well. For the first time we solve self-consistently the mean-field equations for the superconducting gap and the spontaneous magnetization. We discuss the physical features which are different in this model and the standard BCS model and consider their experimental consequences.