Integrating fNIRS and machine learning: shedding light on Parkinson's disease detection.

The purpose of this research is to introduce an approach to assist the diagnosis of Parkinson's disease (PD) by classifying functional near-infrared spectroscopy (fNIRS) studies as PD positive or negative. fNIRS is a non-invasive optical signal modality that conveys the brain's hemodynamic response, specifically changes in blood oxygenation in the cerebral cortex; and its potential as a tool to assist PD detection deserves to be explored since it is non-invasive and cost-effective as opposed to other neuroimaging modalities. Besides the integration of fNIRS and machine learning, a contribution of this work is that various approaches were implemented and tested to find the implementation that achieves the highest performance. All the implementations used a logistic regression model for classification. A set of 792 temporal and spectral features were extracted from each participant's fNIRS study. In the two best performing implementations, an ensemble of feature-ranking techniques was used to select a reduced feature subset, which was subsequently reduced with a genetic algorithm. Achieving optimal detection performance, our approach reached 100 % accuracy, precision, and recall, with an F1 score and area under the curve (AUC) of 1, using 14 features. This significantly advances PD diagnosis, highlighting the potential of integrating fNIRS and machine learning for non-invasive PD detection.

Parkinson's disease patients show delayed hemodynamic changes in primary motor cortex in fine motor tasks and decreased resting-state interhemispheric functional connectivity: a functional near-infrared spectroscopy study.

Significance: People with Parkinson's disease (PD) experience changes in fine motor skills, which is viewed as one of the hallmark signs of this disease. Due to its non-invasive nature and portability, functional near-infrared spectroscopy (fNIRS) is a promising tool for assessing changes related to fine motor skills. Aim: We aim to compare activation patterns in the primary motor cortex using fNIRS, comparing volunteers with PD and sex- and age-matched control participants during a fine motor task and walking. Moreover, inter and intrahemispheric functional connectivity (FC) was investigated during the resting state. Approach: We used fNIRS to measure the hemodynamic changes in the primary motor cortex elicited by a finger-tapping task in 20 PD patients and 20 controls matched for age, sex, education, and body mass index. In addition, a two-minute walking task was carried out. Resting-state FC was also assessed. Results: Patients with PD showed delayed hypoactivation in the motor cortex during the fine motor task with the dominant hand and delayed hyperactivation with the non-dominant hand. The findings also revealed significant correlations among various measures of hemodynamic activity in the motor cortex using fNIRS and different cognitive and clinical variables. There were no significant differences between patients with PD and controls during the walking task. However, there were significant differences in interhemispheric connectivity between PD patients and control participants, with a statistically significant decrease in PD patients compared with control participants. Conclusions: Decreased interhemispheric FC and delayed activity in the primary motor cortex elicited by a fine motor task may one day serve as one of the many potential neuroimaging biomarkers for diagnosing PD.

Fatty Acid-Binding Proteins Identification during the Evolution of Metabolic Syndrome: A Raman Spectroscopy-Based Approach.

Excess fat in abdominal deposits is a risk factor for multiple conditions, including metabolic syndrome (MetS); lipid metabolism plays an essential role in these pathologies; fatty acid-binding proteins (FABPs) are dedicated to the cytosolic transport of fat. FABP4, whose primary source is adipose tissue, is released into the circulation, acting as an adipokine, while FABP5 also accompanies the adverse effects of MetS. FABP4 and 5 are potential biomarkers of MetS, but their behavior during syndrome evolution has not been determined. Raman spectroscopy has been applied as an alternative method to disease biomarker detection. In this work, we detected spectral changes related to FABP4 and 5 in the serum at different points of time, using an animal model of a high-fat diet-induced MetS. FABP4 and 5 spectral changes show a contribution during the evolution of MetS, which indicates alteration to a molecular level that predisposes to established MetS. These findings place FABPs as potential biomarkers of MetS and Raman spectroscopy as an alternative method for MetS assessment.

Response to Comment on "Feasibility of Raman spectroscopy as a potential in vivo tool to screen for pre-diabetes and diabetes".

This letter aims to reply to Bratchenko and Bratchenko's comment on our paper "Feasibility of Raman spectroscopy as a potential in vivo tool to screen for pre-diabetes and diabetes." Our paper analyzed the feasibility of using in vivo Raman measurements combined with machine learning techniques to screen diabetic and prediabetic patients. We argued that this approach yields high overall accuracy (94.3%) while retaining a good capacity to distinguish between diabetic (area under the receiver-operating curve [AUC] = 0.86) and control classes (AUC = 0.97) and a moderate performance for the prediabetic class (AUC = 0.76). Bratchenko and Bratchenko's comment focuses on the possible overestimation of the proposed classification models and the absence of information on the age of participants. In this reply, we address their main concerns regarding our previous manuscript.

Feasibility of Raman spectroscopy as a potential in vivo tool to screen for pre-diabetes and diabetes.

In this article, we investigated the feasibility of using Raman spectroscopy and multivariate analysis method to noninvasively screen for prediabetes and diabetes in vivo. Raman measurements were performed on the skin from 56 patients with diabetes, 19 prediabetic patients and 32 healthy volunteers. These spectra were collected along with reference values provided by the standard glycated hemoglobin (HbA1c) assay. A multiclass principal component analysis and support vector machine (PCA-SVM) model was created from the labeled Raman spectra and was validated through a two-layer cross-validation scheme. Classification accuracy of the model was 94.3% with an area under the receiver operating characteristic curve AUC of 0.76 (0.65-0.84) for the prediabetic group, 0.86 (0.71-0.93) for the diabetic group and 0.97(0.93-0.99) for the control group. Our results suggest the feasibility of using Raman spectroscopy for the classification of prediabetes and diabetes in vivo.

Raman Spectroscopy Study of Skin Biopsies from Patients with Parkinson's Disease: Trends in Alpha-Synuclein Aggregation from the Amide I Region.

Parkinson's disease (PD) is one of the most common neurological pathologies with a high prevalence worldwide. PD is characterized by Lewy bodies, whose major component is the aggregates of α-synuclein (αSyn) protein. Interestingly, recent works have demonstrated that skin biopsy studies are a promising diagnostic tool for evaluating α-synucleinopathies. In this sense, this work focuses on the detection of αSyn in skin biopsies employing Raman spectroscopy, using three different approaches: (i) the in vitro Raman spectrum of α-synuclein, (ii) the ex vivo Raman spectra of human skin biopsies from healthy and Parkinson's disease patients, and (iii) theoretical calculations of the Raman spectra obtained from different model αSyn fragments using density functional theory (DFT). Significant differences in the intensity and location of Raman active frequencies in the amide I region were found when comparing healthy and PD subjects related to α-synuclein conformational changes and variations in their aggregation behavior. In samples from healthy patients, we identified well-known Raman peaks at 1655, 1664, and 1680 cm-1 associated with the normal state of the protein. In PD subjects, shifted Raman bands and intensity variations were found at 1650, 1670, and 1687 cm-1 associated with aggregated forms of the protein. DFT calculations reveal that the shape of the amide I Raman peak in model αSyn fragments strongly depends on the degree of aggregation. Sizable frequency shifts and intensity variations are found within the highly relevant 1600-1700 cm-1 domain, revealing the sensitivity of the amide I Raman band to the changes in the local atomic environment. Interestingly, we obtain that the presence of surrounding waters also affects the structure of the amide I band, leading to the appearance of new peaks on the low-frequency side and a notable broadening of the Raman spectra. These results strongly suggest that, through Raman spectroscopy, it is possible to infer the presence of aggregated forms of αSyn in skin biopsies, a result that could have important implications for understanding α-synuclein related diseases.

Computational and Experimental Analysis of Gold Nanorods in Terms of Their Morphology: Spectral Absorption and Local Field Enhancement.

A nanoparticle's shape and size determine its optical properties. Nanorods are nanoparticles that have double absorption bands associated to surface plasmon oscillations along their two main axes. In this work, we analize the optical response of gold nanorods with numerical simulations and spectral absorption measurements to evaluate their local field enhancement-which is key for surface-enhanced Raman spectroscopic (SERS) applications. Our experimental results are in good agreement with finite element method (FEM) simulations for the spectral optical absorption of the nanoparticles. We also observed a strong dependence of the optical properties of gold nanorods on their geometrical dimension and shape. Our numerical simulations helped us reveal the importance of the nanorods' morphology generated during the synthesis stage in the evaluation of absorption and local field enhancement. The application of these gold nanorods in surface-enhancement Raman spectroscopy is analyzed numerically, and results in a 5.8×104 amplification factor when comparing the values obtained for the nanorod deposited on a dielectric substrate compared to the nanorod immersed in water.

Raman Spectroscopy for Adipose Tissue Assessment in Rat Models of Obesity and Type 1 Diabetes.

Adipose tissue presents structural and functional changes in obesity and type 1 diabetes mellitus (T1DM). In obesity, the size and number of adipocytes and adipokine secretion increases. In T1DM, a loss of adipose tissue suggests changes in the metabolic activity of this tissue. A significant challenge is to find alternative noninvasive methods to evaluate molecular changes in adipose tissue related to obesity and T1DM. Recently, Raman spectroscopy and chemometrics techniques have emerged as a tool for biological tissue analysis. In this work, we propose the use of Raman spectroscopy to characterize spectral differences in adipose tissue from different rat groups (control, obese, and T1DM). The Raman spectra were analyzed using direct band analysis, ratiometric analysis, and chemometric methods (principal component analysis (PCA) and support vector machines (SVMs)). We found that the Raman spectra of obese rats showed significant spectral differences compared to control and diabetic groups related to fatty acids Raman bands. Also, the obese group has a significant decrease in the degree of unsaturation of lipids. The PCA-SVM models showed classification performance ranging from 71.43% to 71.79% accuracy for brown and white adipose tissue samples, respectively. In conclusion, the results demonstrate that Raman spectroscopy can be used as a nondestructive method to assess adipose tissue according to a metabolic condition.

Use of Raman spectroscopy to screen diabetes mellitus with machine learning tools: reply to comment.

We show the spectra of advanced glycation products in response to recent comments made by Bratchenko et al. Our results suggest that information retrieved by Raman spectroscopy is relevant to screening diabetic patients, however, the comparison carried out in our paper, between ANN and SVM, was not fair, because of the erroneous PCA selection procedure and different sources of variation present in the analysis.

Prediction of epileptic seizures with convolutional neural networks and functional near-infrared spectroscopy signals.

There have been different efforts to predict epileptic seizures and most of them are based on the analysis of electroencephalography (EEG) signals; however, recent publications have suggested that functional Near-Infrared Spectroscopy (fNIRS), a relatively new technique, could be used to predict seizures. The objectives of this research are to show that the application of fNIRS to epileptic seizure detection yields results that are superior to those based on EEG and to demonstrate that the application of deep learning to this problem is suitable given the nature of fNIRS recordings. A Convolutional Neural Network (CNN) is applied to the prediction of epileptic seizures from fNIRS signals, an optical modality for recording brain waves. The implementation of the proposed method is presented in this work. Application of CNN to fNIRS recordings showed an accuracy ranging between 96.9% and 100%, sensitivity between 95.24% and 100%, specificity between 98.57% and 100%, a positive predictive value between 98.52% and 100%, and a negative predictive value between 95.39% and 100%. The most important aspect of this research is the combination of fNIRS signals with the particular CNN algorithm. The fNIRS modality has not been used in epileptic seizure prediction. A CNN is suitable for this application because fNIRS recordings are high dimensional data and they can be modeled as three-dimensional tensors for classification.

Use of Raman spectroscopy in the assessment of skin after CO2 ablative fractional laser surgery on acne scars.

BACKGROUND: Ablative fractional laser surgery is a common technique for treating acne scars. However, an in vivo and noninvasive analysis of the histologic variations between acne skin and the resulting resurfaced skin is needed in order to evaluate the wound healing process of the scars induced by the ablative fractional laser surgery. MATERIALS AND METHODS: Nine patients with acne scars underwent a single treatment with a CO2 ablative fractional laser surgery. Collagen presence on the resurfaced skin was noninvasively assessed by means of Raman spectroscopy and principal component analysis. RESULTS: Principal component analysis shows that all the patients presented a collagen regeneration on the resurfaced skin after the laser treatment. CONCLUSION: Collagen plays a crucial role in the wound healing process. By assessing the collagen presence on the skin, it was possible to quantify the regenerative effects of the ablative fractional laser in a noninvasive way.

An automated method for the evaluation of breast cancer using infrared thermography.

Breast cancer is one of the major causes of death for women. Temperature measurement is advantageous because it is non-invasive, non-destructive, and cost-effective. Temperature measurement through infrared thermography is useful to detect changes in blood perfusion that can occur due to inflammation, angiogenesis, or other pathological causes. In this work, we analyzed 206 thermograms of patients with suspected breast cancer, using a classification method, in which thermal asymmetries were computed, the most vascularized areas of each breast were extracted and compared; then these two metrics were added to yield a thermal score, indicative of thermal anomalies. The classification method based on this thermal score allowed us to obtain the test sensitivity of 100 %, specificity of 68.68 %; a positive predictive value of 11.42 % and negative predictive value of 100 %. These results highlight the potential of thermography imaging as adjunctive tool to mammography in breast cancer screening.

Use of Raman spectroscopy to screen diabetes mellitus with machine learning tools.

Type 2 diabetes mellitus (DM2) is one of the most widely prevalent diseases worldwide and is currently screened by invasive techniques based on enzymatic assays that measure plasma glucose concentration in a laboratory setting. A promising plan of action for screening DM2 is to identify molecular signatures in a non-invasive fashion. This work describes the application of portable Raman spectroscopy coupled with several supervised machine-learning techniques, to discern between diabetic patients and healthy controls (Ctrl), with a high degree of accuracy. Using artificial neural networks (ANN), we accurately discriminated between DM2 and Ctrl groups with 88.9-90.9% accuracy, depending on the sampling site. In order to compare the ANN performance to more traditional methods used in spectroscopy, principal component analysis (PCA) was carried out. A subset of features from PCA was used to generate a support vector machine (SVM) model, albeit with decreased accuracy (76.0-82.5%). The 10-fold cross-validation model was performed to validate both classifiers. This technique is relatively low-cost, harmless, simple and comfortable for the patient, yielding rapid diagnosis. Furthermore, the performance of the ANN-based method was better than the typical performance of the invasive measurement of capillary blood glucose. These characteristics make our method a promising screening tool for identifying DM2 in a non-invasive and automated fashion.

Levocarnitine Decreases Intradialytic Hypotension Episodes: A Randomized Controlled Trial.

Intradialytic hypotension is common complication in stage 5 chronic kidney disease patients on hemodialysis. Incidence ranges from 15 to 30%. These patients have levocarnitine deficiency. A randomized, placebo-controlled quadruple-blinded trial was designed to demonstrate the levocarnitine efficiency on intradialytic hypotension prevention. Patients were randomized into four groups, to receive levocarnitine or placebo. During the intervention period, levocarnitine and placebo was administered 0 and 30 min before each hemodialysis session, respectively. During the trial, 33 patients received 1188 hemodialysis sessions. We identified 239 (21.3%) intradialytic hypotension episodes. The intradialytic hypotension episodes were less frequent in the levocarnitine group (9.3%, 60 IH events) (P < 0.001). Hemodialysis is frequently perplexed by intradialytic hypotension episodes. Levocarnitine supplementation before each hemodialysis session efficiently diminishes the intradialytic hypotension episodes. This is a new application method that must be considered and explored.

Altered Functional Connectivity Following an Inflammatory White Matter Injury in the Newborn Rat: A High Spatial and Temporal Resolution Intrinsic Optical Imaging Study.

Very preterm newborns have an increased risk of developing an inflammatory cerebral white matter injury that may lead to severe neuro-cognitive impairment. In this study we performed functional connectivity (fc) analysis using resting-state optical imaging of intrinsic signals (rs-OIS) to assess the impact of inflammation on resting-state networks (RSN) in a pre-clinical model of perinatal inflammatory brain injury. Lipopolysaccharide (LPS) or saline injections were administered in postnatal day (P3) rat pups and optical imaging of intrinsic signals were obtained 3 weeks later. (rs-OIS) fc seed-based analysis including spatial extent were performed. A support vector machine (SVM) was then used to classify rat pups in two categories using fc measures and an artificial neural network (ANN) was implemented to predict lesion size from those same fc measures. A significant decrease in the spatial extent of fc statistical maps was observed in the injured group, across contrasts and seeds (*p = 0.0452 for HbO2 and **p = 0.0036 for HbR). Both machine learning techniques were applied successfully, yielding 92% accuracy in group classification and a significant correlation r = 0.9431 in fractional lesion volume prediction (**p = 0.0020). Our results suggest that fc is altered in the injured newborn brain, showing the long-standing effect of inflammation.

Functional electrical stimulation post-spinal cord injury improves locomotion and increases afferent input into the central nervous system in rats.

BACKGROUND: Functional electrical stimulation (FES) has been found to be effective in restoring voluntary functions after spinal cord injury (SCI) and stroke. However, the central nervous system (CNS) changes that occur in as a result of this therapy are largely unknown. OBJECTIVE: To examine the effects of FES on the restoration of voluntary locomotor function of the CNS in a SCI rat model. METHODS: SCI rats were instrumented with chronic FES electrodes in the hindlimb muscles and were divided into two groups: (a) FES therapy and (b) sedentary. At day 7 post-SCI, the animals were assessed for locomotion performance by using a Basso, Beattie and Bresnahan (BBB) scale. They were then anesthetized for a terminal in vivo experiment. The lumbar spinal cord and somatosensory cortex were exposed and the instrumented muscles were stimulated electrically. Associated neurovascular responses in the CNS were recorded with an intrinsic optical imaging system. RESULTS: FES greatly improved locomotion recovery by day 7 post-SCI, as measured by BBB scores (P < 0.05): (a) FES 10 ± 2 and (b) controls 3 ± 1. Furthermore, the FES group showed a significant increase (P < 0.05) of neurovascular activation in the spinal cord and somatosensory cortex when the muscles were stimulated between 1 and 3 motor threshold (MT). CONCLUSION: Hind limb rehabilitation with FES is an effective strategy to improve locomotion during the acute phase post-SCI. The results of this study indicate that after FES, the CNS preserves/acquires the capacity to respond to peripheral electrical stimulation.

Optical imaging of resting-state functional connectivity in a novel arterial stiffness model.

This study aims to assess the impact of unilateral increases in carotid stiffness on cortical functional connectivity measures in the resting state. Using a novel animal model of induced arterial stiffness combined with optical intrinsic signals and laser speckle imaging, resting state functional networks derived from hemodynamic signals are investigated for their modulation by isolated changes in stiffness of the right common carotid artery. By means of seed-based analysis, results showed a decreasing trend of homologous correlation in the motor and cingulate cortices. Furthermore, a graph analysis indicated a randomization of the cortex functional networks, suggesting a loss of connectivity, more specifically in the motor cortex lateral to the treated carotid, which however did not translate in differentiated metabolic activity.

Optical imaging of acute epileptic networks in mice.

The potential of intrinsic optical imaging and resting-state analysis under anesthetized conditions as a tool to study brain networks associated with epileptic seizures is investigated. Using an acute model of epileptiform activity, the 4-aminopyridine model in live mice, we observe the changes in resting-state networks with the onset of seizure activity and in conditions of spiking activity. Resting-state networks identified before and after the onset of epileptiform activity show both decreased and increased homologous correlations, with a small dependence on seizure intensity. The observed changes are not uniform across the different hemodynamic measures, suggesting a potential decoupling between blood flow and metabolism in the low-frequency networks. This study supports the need for a more extensive investigation of epileptic networks including more than one independent hemodynamic measurement.

Carotid calcification in mice: a new model to study the effects of arterial stiffness on the brain.

BACKGROUND: Arterial stiffness has been identified as an important risk factor for cognitive decline. However, its effects on the brain's health are unknown, and there is no animal model available to study the precise impact of arterial stiffness on the brain. Therefore, the objective of the study was to develop and characterize a new model specific to arterial stiffness in order to study its effects on the brain. METHODS AND RESULTS: Calcium chloride (CaCl2) was applied to carotid arteries of mice, inducing an increase in collagen distribution and intima-media thickness, a fragmentation of elastin, a decrease in arterial compliance and distensibility, and an increase in cerebral blood flow pulsatility (n=3 to 11). Calcium deposits were only present at the site of CaCl2 application, and there was no increase in systemic blood pressure or change in vessel radius making this model specific for arterial stiffness. The effects of carotid stiffness were then assessed in the brain. Carotid calcification induced an increase in the production of cerebral superoxide anion and neurodegeneration, detected with Fluoro-Jade B staining, in the hippocampus (n=3 to 5), a key region for memory and cognition. CONCLUSIONS: A new model of arterial stiffness based on carotid calcification was developed and characterized. This new model meets all the characteristics of arterial stiffness, and its specificity allows the study of the effects of arterial stiffness on the brain.

Imaging of an inflammatory injury in the newborn rat brain with photoacoustic tomography.

BACKGROUND: The precise assessment of cerebral saturation changes during an inflammatory injury in the developing brain, such as seen in periventricular leukomalacia, is not well defined. This study investigated the impact of inflammation on locoregional cerebral oxygen saturation in a newborn rodent model using photoacoustic imaging. METHODS: 1 mg/kg of lipopolysaccharide(LPS) diluted in saline or saline alone was injected under ultrasound guidance directly in the corpus callosum of P3 rat pups. Coronal photoacoustic images were carried out 24 h after LPS exposure. Locoregional oxygen saturation (SO2) and resting state connectivity were assessed in the cortex and the corpus callosum. Microvasculature was then evaluated on cryosection slices by lectin histochemistry. RESULTS: Significant reduction of SO2 was found in the corpus callosum; reduced SO2 was also found in the cortex ipsilateral to the injection site. Seed-based functional connectivity analysis showed that bilateral connectivity was not affected by LPS exposure. Changes in locoregional oxygen saturation were accompanied by a significant reduction in the average length of microvessels in the left cortex but no differences were observed in the corpus callosum. CONCLUSION: Inflammation in the developing brain induces marked reduction of locoregional oxygen saturation, predominantly in the white matter not explained by microvascular degeneration. The ability to examine regional saturation offers a new way to monitor injury and understand physiological disturbance non-invasively.