Retrospective Continuous-Time Blood Glucose Estimation in Free Living Conditions with a Non-Invasive Multisensor Device.

Even if still at an early stage of development, non-invasive continuous glucose monitoring (NI-CGM) sensors represent a promising technology for optimizing diabetes therapy. Recent studies showed that the Multisensor provides useful information about glucose dynamics with a mean absolute relative difference (MARD) of 35.4% in a fully prospective setting. Here we propose a method that, exploiting the same Multisensor measurements, but in a retrospective setting, achieves a much better accuracy. Data acquired by the Multisensor during a long-term study are retrospectively processed following a two-step procedure. First, the raw data are transformed to a blood glucose (BG) estimate by a multiple linear regression model. Then, an enhancing module is applied in cascade to the regression model to improve the accuracy of the glucose estimation by retrofitting available BG references through a time-varying linear model. MARD between the retrospectively reconstructed BG time-series and reference values is 20%. Here, 94% of values fall in zone A or B of the Clarke Error Grid. The proposed algorithm achieved a level of accuracy that could make this device a potential complementary tool for diabetes management and also for guiding prediabetic or nondiabetic users through life-style changes.

Non-invasive continuous glucose monitoring with multi-sensor systems: a Monte Carlo-based methodology for assessing calibration robustness.

In diabetes research, non-invasive continuous glucose monitoring (NI-CGM) devices represent a new and appealing frontier. In the last years, some multi-sensor devices for NI-CGM have been proposed, which exploit several sensors measuring phenomena of different nature, not only for measuring glucose related signals, but also signals reflecting some possible perturbing processes (temperature, blood perfusion). Estimation of glucose levels is then obtained combining these signals through a mathematical model which requires an initial calibration step exploiting one reference blood glucose (RBG) sample. Even if promising results have been obtained, especially in hospitalized volunteers, at present the temporal accuracy of NI-CGM sensors may suffer because of environmental and physiological interferences. The aim of this work is to develop a general methodology, based on Monte Carlo (MC) simulation, to assess the robustness of the calibration step used by NI-CGM devices against these disturbances. The proposed methodology is illustrated considering two examples: the first concerns the possible detrimental influence of sweat events, while the second deals with calibration scheduling. For implementing both examples, 45 datasets collected by the Solianis Multisensor system are considered. In the first example, the MC methodology suggests that no further calibration adjustments are needed after the occurrence of sweat events, because the "Multisensor+model" system is able to deal with the disturbance. The second case study shows how to identify the best time interval to update the model's calibration for improving the accuracy of the estimated glucose. The methodology proposed in this work is of general applicability and can be helpful in making those incremental steps in NI-CGM devices development needed to further improve their performance.

Physiological sensor data cleaning with autoencoders.

Objective.Physiological sensor data (e.g. photoplethysmograph) is important for remotely monitoring patients' vital signals, but is often affected by measurement noise. Existing feature-based models for signal cleaning can be limited as they might not capture the full signal characteristics.Approach.In this work we present a deep learning framework for sensor signal cleaning based on dilated convolutions which capture the coarse- and fine-grained structure in order to classify whether a signal is noisy or clean. However, since obtaining annotated physiological data is costly and time-consuming we propose an autoencoder-based semi-supervised model which is able to learn a representation of the sensor signal characteristics, also adding an element of interpretability.Main results.Our proposed models are over 8% more accurate than existing feature-based approaches with half the false positive/negative rates. Finally, we show that with careful tuning (that can be improved further), the semi-supervised model outperforms supervised approaches suggesting that incorporating the large amounts of available unlabeled data can be advantageous for achieving high accuracy (over 90%) and minimizing the false positive/negative rates.Significance.Our approach enables us to reliably separate clean from noisy physiological sensor signal that can pave the development of reliable features and eventually support decisions regarding drug efficacy in clinical trials.

An observational study to assess validity and reliability of smartphone sensor-based gait and balance assessments in multiple sclerosis: Floodlight GaitLab protocol.

Background: Gait and balance impairments are often present in people with multiple sclerosis (PwMS) and have a significant impact on quality of life and independence. Gold-standard quantitative tools for assessing gait and balance such as motion capture systems and force plates usually require complex technical setups. Wearable sensors, including those integrated into smartphones, offer a more frequent, convenient, and minimally burdensome assessment of functional disability in a home environment. We developed a novel smartphone sensor-based application (Floodlight) that is being used in multiple research and clinical contexts, but a complete validation of this technology is still lacking. Methods: This protocol describes an observational study designed to evaluate the analytical and clinical validity of Floodlight gait and balance tests. Approximately 100 PwMS and 35 healthy controls will perform multiple gait and balance tasks in both laboratory-based and real-world environments in order to explore the following properties: (a) concurrent validity of the Floodlight gait and balance tests against gold-standard assessments; (b) reliability of Floodlight digital measures derived under different controlled gait and balance conditions, and different on-body sensor locations; (c) ecological validity of the tests; and (d) construct validity compared with clinician- and patient-reported assessments. Conclusions: The Floodlight GaitLab study (ISRCTN15993728) represents a critical step in the technical validation of Floodlight technology to measure gait and balance in PwMS, and will also allow the development of new test designs and algorithms.

Italian contributions to the development of continuous glucose monitoring sensors for diabetes management.

Monitoring glucose concentration in the blood is essential in the therapy of diabetes, a pathology which affects about 350 million people around the World (three million in Italy), causes more than four million deaths per year and consumes a significant portion of the budget of national health systems (10% in Italy). In the last 15 years, several sensors with different degree of invasiveness have been proposed to monitor glycemia in a quasi-continuous way (up to 1 sample/min rate) for relatively long intervals (up to 7 consecutive days). These continuous glucose monitoring (CGM) sensors have opened new scenarios to assess, off-line, the effectiveness of individual patient therapeutic plans from the retrospective analysis of glucose time-series, but have also stimulated the development of innovative on-line applications, such as hypo/hyper-glycemia alert systems and artificial pancreas closed-loop control algorithms. In this review, we illustrate some significant Italian contributions, both from industry and academia, to the growth of the CGM sensors research area. In particular, technological, algorithmic and clinical developments performed in Italy will be discussed and put in relation with the advances obtained in the field in the wider international research community.

Simulating the impact of noise on gait features extracted from smartphone sensor-data for the remote assessment of movement disorders.

Signs and symptoms of movement disorders can be remotely measured at home through sensor-based assessment of gait. However, sensor noise may impact the robustness of such assessments, in particular in a Bring-Your-Own-Device setting where the quality of sensors might vary. Here, we propose a framework to study the impact of inertial measurement unit noise on sensor-based gait features. This framework includes synthesizing realistic acceleration signals from the lower back during a gait cycle in OpenSim, estimating the magnitude of sensor noise from five smartphone models, perturbing the synthesized acceleration signal with the estimated noise in a Monte Carlo simulation, and computing gait features. In addition, we show that realistic levels of sensor noise have only a negligible impact on step power, a measure of gait.

A quality metric for heart rate variability from photoplethysmogram sensor data.

Heart rate variability (HRV) measures the regularity between consecutive heartbeats driven by the balance between the sympathetic and parasympathetic branches of the autonomous nervous system. Wearable devices embedding photoplethysmogram (PPG) technology can be used to derive HRV, creating many opportunities for remote monitoring of this physiological parameter. However, uncontrolled conditions met in daily life pose several challenges related to disturbances that can deteriorate the PPG signal, making the calculation of HRV metrics untrustworthy and not reliable. In this work, we propose a HRV quality metric that is directly related to the HRV accuracy and can be used to distinguish between accurate and inaccurate HRV values. A parametric supervised approach estimates HRV accuracy using a model whose inputs are features extracted from the PPG signal and the output is the HRV error between HRV metrics obtained from the PPG and the ECG collected during an experimental protocol involving several activities. The estimated HRV accuracy of the model is used as an indication of the HRV quality.

First Experiences With a Wearable Multisensor in an Outpatient Glucose Monitoring Study, Part I: The Users' View.

BACKGROUND: Extensive past work showed that noninvasive continuous glucose monitoring with a wearable Multisensor device worn on the upper arm provides useful information about glucose trends to improve diabetes therapy in controlled and semicontrolled conditions. METHODS: To test previous findings also in uncontrolled in-clinic and outpatient conditions, a long-term study has been conducted to collect Multisensor and reference glucose data in a population of 20 type 1 diabetes subjects. A total of 1072 study days were collected and a fully on-line compatible algorithmic routine linking Multisensor data to glucose applied to estimate glucose trends noninvasively. The operation of a digital log book, daily semiautomated data transfer and at least 10 daily SMBG values were requested from the patient. RESULTS: Results showed that the Multisensor is capable of indicating glucose trends. It can do so in 9 out of 10 cases either correctly or with one level of discrepancy. This means that in 90% of all cases the Multisensor shows the glucose dynamic to rapidly increase or at least increase. CONCLUSIONS: The Multisensor and the algorithmic routine used in controlled conditions can track glucose trends in all patients, also in uncontrolled conditions. Training of the patient proved to be essential. The workload imposed on patients was significant and should be reduced in the next step with further automation. The feature of glucose trend indication was welcomed and very much appreciated by patients; this value creation makes a strong case for the justification of wearing a wearable.

First Experiences With a Wearable Multisensor Device in a Noninvasive Continuous Glucose Monitoring Study at Home, Part II: The Investigators' View.

BACKGROUND: Extensive past work showed that noninvasive continuous glucose monitoring with a wearable multisensor device worn on the upper arm provides useful information about glucose trends to improve diabetes therapy in controlled and semicontrolled conditions. METHOD: To test previous findings also in uncontrolled conditions, a long term at home study has been organized to collect multisensor and reference glucose data in a population of 20 type 1 diabetes subjects. A total of 1072 study days were collected and a fully on-line compatible algorithmic routine linking multisensor data to glucose applied to estimate glucose levels noninvasively. RESULTS: The algorithm used here calculates glucose values from sensor data and adds a constant obtained by a daily calibration. It provides point inaccuracy measured by a MARD of 35.4 mg/dL on test data. This is higher than current state-of-the-art minimally invasive devices, but still 86.9% of glucose rate points fall within the zone AR+BR. CONCLUSIONS: The multisensor device and the algorithmic routine used earlier in controlled conditions tracks glucose changes also in uncontrolled conditions, although with lower accuracy. The examination of learning curves suggests that obtaining more data would not improve the results. Therefore, further efforts would focus on the development of more complex algorithmic routines able to compensate for environmental and physiological confounders better.

EEG signal features extraction based on fractal dimension.

The spread of electroencephalography (EEG) in countless applications has fostered the development of new techniques for extracting synthetic and informative features from EEG signals. However, the definition of an effective feature set depends on the specific problem to be addressed and is currently an active field of research. In this work, we investigated the application of features based on fractal dimension to a problem of sleep identification from EEG data. We demonstrated that features based on fractal dimension, including two novel indices defined in this work, add valuable information to standard EEG features and significantly improve sleep identification performance.

The Effect of a Global, Subject, and Device-Specific Model on a Noninvasive Glucose Monitoring Multisensor System.

BACKGROUND: We study here the influence of different patients and the influence of different devices with the same patients on the signals and modeling of data from measurements from a noninvasive Multisensor glucose monitoring system in patients with type 1 diabetes. The Multisensor includes several sensors for biophysical monitoring of skin and underlying tissue integrated on a single substrate. METHOD: Two Multisensors were worn simultaneously, 1 on the upper left and 1 on the upper right arm by 4 patients during 16 study visits. Glucose was administered orally to induce 2 consecutive hyperglycemic excursions. For the analysis, global (valid for a population of patients), personal (tailored to a specific patient), and device-specific multiple linear regression models were derived. RESULTS: We find that adjustments of the model to the patients improves the performance of the glucose estimation with an MARD of 17.8% for personalized model versus a MARD of 21.1% for the global model. At the same time the effect of the measurement side is negligible. The device can equally well measure on the left or right arm. We also see that devices are equal in the linear modeling. Thus hardware calibration of the sensors is seen to be sufficient to eliminate interdevice differences in the measured signals. CONCLUSIONS: We demonstrate that the hardware of the 2 devices worn on the left and right arms are consistent yielding similar measured signals and thus glucose estimation results with a global model. The 2 devices also return similar values of glucose errors. These errors are mainly due to nonstationarities in the measured signals that are not solved by the linear model, thus suggesting for more sophisticated modeling approaches.

Glucose variability indices in type 1 diabetes: parsimonious set of indices revealed by sparse principal component analysis.

BACKGROUND: Continuous glucose monitoring (CGM) time-series are often analyzed, retrospectively, to investigate glucose variability (GV), a risk factor for the development of complications in type 1 diabetes (T1D). In the literature, several tens of different indices for GV quantification have been proposed, but many of them carry very similar information. The aim of this article is to select a relatively small subset of GV indices from a wider pool of metrics, to obtain a parsimonious but still comprehensive description of GV in T1D datasets. MATERIALS AND METHODS: A pool of 25 GV indices was evaluated on two CGM time-series datasets of 17 and 16 T1D subjects, respectively, collected during the European Union Seventh Framework Programme project "Diadvisor" (2008-2012) in two different clinical research centers using the Dexcom(®) (San Diego, CA) SEVEN(®) Plus. After the indices were centered and scaled, the Sparse Principal Component Analysis (SPCA) technique was used to determine a reduced set of metrics that allows preserving a high percentage of the variance of the whole original set. In order to assess whether or not the selected subset of GV indices is dataset-dependent, the analysis was applied to both datasets, as well as to the one obtained by merging them. RESULTS: SPCA revealed that a subset of up to 10 different GV indices can be sufficient to preserve more than the 60% of the variance originally explained by all the 25 variables. It is remarkable that four of these GV indices (i.e., Index of Glycemic Control, percentage of Glycemic Risk Assessment Diabetes Equation score due to euglycemia, percentage Coefficient of Variation, and Low Blood Glucose Index) were selected for all the considered T1D datasets. CONCLUSIONS: The SPCA methodology appears a suitable candidate to identify, among the large number of literature GV indices, subsets that allow obtaining a parsimonious, but still comprehensive, description of GV.

Non-invasive continuous glucose monitoring: improved accuracy of point and trend estimates of the Multisensor system.

Non-invasive continuous glucose monitoring (NI-CGM) sensors are still at an early stage of development, but, in the near future, they could become particularly appealing in diabetes management. Solianis Monitoring AG (Zurich, Switzerland) has proposed an approach for NI-CGM based on a multi-sensor concept, embedding primarily dielectric spectroscopy and optical sensors. This concept requires a mathematical model able to estimate glucose levels from the 150 channels directly measured through the Multisensor. A static multivariate linear regression model (with order and parameters common to the entire population of subjects) was proposed for such a scope (Caduff et al., Biosens Bioelectron 26:3794-3800, 2011). The aim of this work is to evaluate the accuracy in the estimation of glucose levels and trends that the NI-CGM Multisensor platform can achieve by exploiting different techniques for model identification, namely, ordinary least squares, subset variable selection, partial least squares and least absolute shrinkage and selection operator (LASSO). Data collected in human beings monitored for a total of 45 study days were used for model identification and model test. Several metrics of standard use in the diabetes scientific community to measure point and clinical accuracy of glucose sensors were used to assess the models. Results indicate that the LASSO technique is superior to the others shrinking many channel weights to zero thus leading to smoother glucose profiles and resulting in a more robust model to possible artifacts in the Multisensor data. Although, as expected, the performance of the NI-CGM system with the LASSO model is not yet comparable with that of enzyme-based needle glucose sensors, glucose trends are satisfactorily estimated. Considering the non-invasive nature of the multi-sensor platform, this result can have an immediate impact in the current clinical practice, e.g., to integrate sparse self-monitoring of blood glucose data with an indication of the glucose trend to aid the diabetic patient in dealing with, or even preventing in the short time scale, the threats of critical events such as hypoglycaemia.

Assessment of linear regression techniques for modeling multisensor data for non-invasive continuous glucose monitoring.

New scenarios in diabetes treatment have been opened in the last ten years by continuous glucose monitoring (CGM) sensors. In particular, Non-Invasive CGM sensors are particularly appealing, even though they are still at an early stage of development. Solianis Monitoring AG (Zürich, Switzerland) has proposed an approach based on a multisensor concept, embedding primarily dielectric spectroscopy and optical sensors. This concept requires a mathematical model able to reconstruct the glucose concentration from the 150 channels measured with the device. Assuming a multivariate linear regression model (valid and usable for different individuals), the aim of this paper is the assessment of some techniques usable for determining such a model, namely Ordinary Least Squares (OLS), Partial Least Squares (PLS) and Least Absolute Shrinkage and Selection Operator (LASSO). Once the model is identified on a training set, the accuracy of prospective glucose profiles estimated from "unseen" multisensor data is assessed. Preliminary results obtained from 18 in-clinic study days show that sufficiently accurate reconstruction of glucose levels can be achieved if suitable model identification techniques, such as LASSO, are considered.