Continuous Glucose, Insulin and Lifestyle Data Augmentation in Artificial Pancreas Using Adaptive Generative and Discriminative Models.

Artificial pancreas requires data from multiple sources for accurate insulin dose estimation. These include data from continuous glucose sensors, past insulin dosage information, meal quantity and time and physical activity data. The effectiveness of closed-loop diabetes management systems might be hampered by the absence of these data caused by device error or lack of compliance by patients. In this study, we demonstrate the effect of output sequence length-driven generative and discriminative model selection in high quality data generation and augmentation. This novel generative adversarial network (GAN) based architecture automatically selects the generator and discriminator architecture based on the desired output sequence length. The proposed model is able to generate glucose, physical activity, meal information data for individual patients. The discriminative scores for Ohio T1DM (2018) dataset were 0.17 ±0.03 (Inputs: CGM, CHO, Insulin) and 0.15 ±0.02 (Inputs: CGM, CHO, Insulin, Heart Rate, Steps) and for Ohio T1D (2020) dataset was 0.16 ±0.02 (Inputs: CGM, CHO, Insulin) and 0.15 ±0.02 (Inputs: CGM, CHO, Insulin, acceleration). A mixture of generated and real data was used to test predictive scores for glucose forecasting models. The best RMSE and MARD achieved for OhioT1DM patients were 17.19 ±3.22 and 7.14 ±1.76 for PH=30 min with CGM, CHO, Insulin, heartrate and steps as inputs. Similarly, the RMSE and MARD for real+synthetic data were 15.63 ±2.57 and 5.86 ±1.69 respectively. Compared to existing generative models, we demonstrate that sequence length based architecture selection leads to better synthetic data generation for multiple output sequences (CGM, CHO, Insulin) and forecasting accuracy.

InsNET: Accurate Basal and Bolus Insulin Dose Prediction for Closed Loop Diabetes Management.

It has been demonstrated that closed-loop diabetes management results in better glycemic control and greater compliance than open-loop diabetes management. Deep learning models have been used to implement different components of artifical pancreas. In this work, a novel deep learning model InsNET has been proposed to estimate the basal and bolus insulin level and insulin bolus in patients with type I diabetes utilizing subcutaneous insulin infusion pumps for closed loop diabetes management system. The proposed InsNET is formed with a Wide-Deep combination of LSTM and GRU layers. Additionally, physical activity level has been included as an input in comparison to previous models where only past glucose levels (CGM), meal intake (CHO) and past insulin dosage were used as inputs. The proposed model was tested on In-silico data, and it achieved a Mean Absolute Error (MAE) of 0.002 and Root Mean Squared Error (RMSE) of 0.007 for UVA/Padova Dataset and MAE of 0.001 and RMSE OF 0.003 for mGIPsim Dataset.Clinical relevance- Insulin dose determination is an important as aspect of artificial pancreas. This work describes a deep learning model to determine accurate basal and bolus insulin dosage.

Glucose Prediction using Wide-Deep LSTM Network for Accurate Insulin Dosing in Artificial Pancreas.

Closed-loop diabetes management has been shown to indicate improved glycaemic control and better compliance over open loop diabetes management. Currently, commercially available diabetes management devices rely on continuous glucose monitoring (CGM) sensors which monitor glucose levels from the interstitial fluid (ISF). As there exists a physiological delay between the blood glucose levels compared to the ISF glucose levels, it is crucial to predict or forecast glucose levels, in order to prevent hyperglycaemic events due to delayed insulin dosing. Accuracy of the forecast also needs to be optimum such that overdosing on insulin does not lead to hypoglycaemia. In this paper, we describe a novel Long Short Term Memory (LSTM) network which follows a wide and deep approach for different features to deliver an accurate glucose prediction output. It achieved a Mean Absolute Relative Difference (MARD) of 2.61 and Root Mean Squared Error (MSE) of 5.04. Clinical relevance- This work is relevant for closed-loop diabetes management devices, which are currently being used to manage Type 1 Diabetes (T1D).

Aptasensor for Quantification of Leptin Through PCR Amplification of Short DNA-Aptamers.

Protein quantification is traditionally performed through enzyme-linked immunosorbent assay (ELISA), which involves long preparation times. To overcome this, new approaches use aptamers as an alternative to antibodies. In this paper, we present a new approach to quantify proteins with short DNA aptamers through polymerase chain reaction (PCR) resulting in shorter protocol times with comparatively improved limits of detection. The proposed method includes a novel way to quantify both the target protein and the corresponding short DNA-aptamers simultaneously, which also allows us to fully characterize the performance of aptasensors. Human leptin is used as a target protein to validate this technique, because it is considered an important biomarker for obesity-related studies. In our experiments, we achieved the lowest limit of detection of 100 pg/mL within less than 2 h, a limit affected by the dissociation constant of the leptin aptamer, which could be improved by selecting a more specific aptamer. Because of the simple and inexpensive approach, this technique can be employed for Lab-On-Chip implementations and for rapid "on-site" quantification of proteins.

Closed-Loop Implantable Therapeutic Neuromodulation Systems Based on Neurochemical Monitoring.

Closed-loop or intelligent neuromodulation allows adjustable, personalized neuromodulation which usually incorporates the recording of a biomarker, followed by implementation of an algorithm which decides the timing (when?) and strength (how much?) of stimulation. Closed-loop neuromodulation has been shown to have greater benefits compared to open-loop neuromodulation, particularly for therapeutic applications such as pharmacoresistant epilepsy, movement disorders and potentially for psychological disorders such as depression or drug addiction. However, an important aspect of the technique is selection of an appropriate, preferably neural biomarker. Neurochemical sensing can provide high resolution biomarker monitoring for various neurological disorders as well as offer deeper insight into neurological mechanisms. The chemicals of interest being measured, could be ions such as potassium (K+), sodium (Na+), calcium (Ca2+), chloride (Cl-), hydrogen (H+) or neurotransmitters such as dopamine, serotonin and glutamate. This review focusses on the different building blocks necessary for a neurochemical, closed-loop neuromodulation system including biomarkers, sensors and data processing algorithms. Furthermore, it also highlights the merits and drawbacks of using this biomarker modality.

Influence of Cholecystokinin-8 on Compound Nerve Action Potentials from Ventral Gastric Vagus in Rats.

OBJECTIVE: Vagus Nerve Stimulation (VNS) has shown great promise as a potential therapy for a number of conditions, such as epilepsy, depression and for Neurometabolic Therapies, especially for treating obesity. The objective of this study was to characterize the left ventral subdiaphragmatic gastric trunk of vagus nerve (SubDiaGVN) and to analyze the influence of intravenous injection of gut hormone cholecystokinin octapeptide (CCK-8) on compound nerve action potential (CNAP) observed on the same branch, with the aim of understanding the impact of hormones on VNS and incorporating the methods and results into closed loop implant design. METHODS: The cervical region of the left vagus nerve (CerVN) of male Wistar rats was stimulated with electric current and the elicited CNAPs were recorded on the SubDiaGVN under four different conditions: Control (no injection), Saline, CCK1 (100[Formula: see text]pmol/kg) and CCK2 (1000[Formula: see text]pmol/kg) injections. RESULTS: We identified the presence of A[Formula: see text], B, C1, C2, C3 and C4 fibers with their respective velocity ranges. Intravenous administration of CCK in vivo results in selective, statistically significant reduction of CNAP components originating from A and B fibers, but with no discernible effect on the C fibers in [Formula: see text] animals. The affected CNAP components exhibit statistically significant ([Formula: see text] and [Formula: see text]) higher normalized stimulation thresholds. CONCLUSION: This approach of characterizing the vagus nerve can be used in closed loop systems to determine when to initiate VNS and also to tune the stimulation dose, which is patient-specific and changes over time.

Extracellular pH monitoring for use in closed-loop vagus nerve stimulation.

OBJECTIVE: Vagal nerve stimulation (VNS) has shown potential benefits for obesity treatment; however, current devices lack physiological feedback, which limit their efficacy. Changes in extracellular pH (pHe) have shown to be correlated with neural activity, but have traditionally been measured with glass microelectrodes, which limit their in vivo applicability. APPROACH: Iridium oxide has previously been shown to be sensitive to fluctuations in pH and is biocompatible. Iridium oxide microelectrodes were inserted into the subdiaphragmatic vagus nerve of anaesthetised rats. Introduction of the gut hormone cholecystokinin (CCK) or distension of the stomach was used to elicit vagal nerve activity. MAIN RESULTS: Iridium oxide microelectrodes have sufficient pH sensitivity to readily detect changes in pHe associated with both CCK and gastric distension. Furthermore, a custom-made Matlab script was able to use these changes in pHe to automatically trigger an implanted VNS device. SIGNIFICANCE: This is the first study to show pHe changes in peripheral nerves in vivo. In addition, the demonstration that iridium oxide microelectrodes are sufficiently pH sensitive as to measure changes in pHe associated with physiological stimuli means they have the potential to be integrated into closed-loop neurostimulating devices.

Injection moulded microneedle sensor for real-time wireless pH monitoring.

This paper describes the development of an array of individually addressable pH sensitive microneedles using injection moulding and their integration within a portable device for real-time wireless recording of pH distributions in biological samples. The fabricated microneedles are subjected to gold patterning followed by electrodeposition of iridium oxide to sensitize them to 0.07 units of pH change. Miniaturised electronics suitable for the sensors readout, analog-to-digital conversion and wireless transmission of the potentiometric data are embodied within the device, enabling it to measure real-time pH of soft biological samples such as muscles. In this paper, real-time recording of the cardiac pH distribution, during ischemia followed by reperfusion cycles in cardiac muscles of male Wistar rats has been demonstrated by using the microneedle array.

Gingival fine needle aspiration cytology in acute leukemia.

BACKGROUND: Oral findings in acute leukemia (AL) are common and could be the presenting feature of the disease, namely, gingival enlargement, ulceration, bleeding, and infection. Gingival enlargement in AL is either due to leukemic infiltration, or due to reactive hyperplasia. To differentiate between them a biopsy is required, but being highly contraindicated, biopsy has been substituted in this study by fine needle aspiration cytology (FNAC). METHOD: Gingival FNAC was performed on different sites in the upper and lower gingiva. Each site represents an interdental papilla, which was selected according to the clinical presentation, i.e., being enlarged or not. Seventy-two adult AL patients received a cytological and clinical examination in this study, and the cases were classified and categorized according to the French-American-British (FAB) criteria. RESULTS: Twenty-one cases were diagnosed as being infiltrated, 16 with gingival enlargement, 4 with no evidence of enlargement. In one case the infiltration affected the alveolar mucosa of an edentulous patient. In six cases the ginigva was enlarged without being infiltrated (reactive hyperplasia). Leukemic gingival enlargement was seen mostly in patients with acute myeloid leukemia, particularly M4 and M5 subtypes; however, two patients with acute mixed lineage leukemia (AMLL) were both affected with leukemic infiltration. CONCLUSION: FNAC was shown to be a simple, non-traumatic and useful diagnostic procedure for screening leukemic infiltration in gingival tissues in AL patients.