Search

This work addresses the employment of Machine Learning (ML) and Domain Adaptation (DA) in the framework of Brain-Computer Interfaces (BCIs) based on Steady-State Visually Evoked Potentials (SSVEPs). Currently, all the state-of-the-art classification strategies do not consider the high non-stationarity typical of brain signals. This can lead to poor performance, expecially when short-time signals have to be considered to allow real-time human-environment interaction. In this regard, ML and DA techniques can represent a suitable strategy to enhance the performance of SSVEPs classification pipelines. In particular, the employment of a two-step DA technique is proposed: first, the standardization of the data per subject is performed by exploiting a part of unlabeled test data during the training stage; second, a similarity measure between subjects is considered in the selection of the validation sets. The proposal was applied to three classifiers to verify the statistical significance of the improvements over the standard approaches. These classifiers were validated and comparatively tested on a well-known public benchmark dataset. An appropriate validation method was used in order to simulate real-world usage. The experimental results show that the proposed approach significantly improves the classification accuracy of SSVEPs. In fact, up to 62.27 % accuracy was achieved also in the case of short-time signals (i.e., 1.0 s). This represents a further confirmation of the suitability of advanced ML to improve the performance of BCIs for daily-life applications.

An innovative algorithm to automatically assess blood perfusion quality of the intestinal sector in laparoscopic colorectal surgery is proposed. Traditionally, the uniformity of the brightness in indocyanine green-based fluorescence consists only in a qualitative, empirical evaluation, which heavily relies on the surgeon’s subjective assessment. As such, this leads to assessments that are strongly experience-dependent. To overcome this limitation, the proposed algorithm assesses the level and uniformity of indocyanine green used during laparoscopic surgery. The algorithm adopts a Feed Forward Neural Network receiving as input a feature vector based on the histogram of the green band of the input image. It is used to (i) acquire information related to perfusion during laparoscopic colorectal surgery, and (ii) support the surgeon in assessing objectively the outcome of the procedure. In particular, the algorithm provides an output that classifies the perfusion as adequate or inadequate. The algorithm was validated on videos captured during surgical procedures carried out at the University Hospital Federico II in Naples, Italy. The obtained results show a classification accuracy equal to $$99.9\%$$ 99.9 % , with a repeatability of $$1.9\%$$ 1.9 % . Finally, the real-time operation of the proposed algorithm was tested by analyzing the video streaming captured directly from an endoscope available in the OR.

An instrument for remote control of the robot by wearable brain–computer interface (BCI) is proposed for rehabilitating children with attention-deficit/hyperactivity disorder (ADHD). Augmented reality (AR) glasses generate flickering stimuli, and a single-channel electroencephalographic BCI detects the elicited steady-state visual evoked potentials (SSVEPs). This allows benefiting from the SSVEP robustness by leaving available the view of robot movements. Together with the lack of training, a single channel maximizes the device’s wearability, fundamental for the acceptance by ADHD children. Effectively controlling the movements of a robot through a new channel enhances rehabilitation engagement and effectiveness. A case study at an accredited rehabilitation center on ten healthy adult subjects highlighted an average accuracy higher than 83%, with information transfer rate (ITR) up to 39 b/min. Preliminary further tests on four ADHD patients between six- and eight-years old provided highly positive feedback on device acceptance and attentional performance.

An extended-reality (XR) platform for real-time monitoring of patients’ health during surgical procedures is proposed. The proposed system provides real-time access to a comprehensive set of patients’ information, which are made promptly available to the surgical team in the operating room (OR). In particular, the XR platform supports the medical staff by automatically acquiring the patient’s vitals from the operating room instrumentation and displaying them in real-time directly on an XR headset. Furthermore, information regarding the patient clinical record is also shown upon request. Finally, the XR-based monitoring platform also allows displaying in XR the video stream coming directly from the endoscope. The innovative aspect of the proposed XR-based monitoring platform lies in the comprehensiveness of the available information, in its modularity and flexibility (in terms of adaption to different sources of data), ease of use, and most importantly, in a reliable communication, which are critical requirements for the healthcare field. To validate the proposed system, experimental tests were conducted using instrumentation typically available in the operating room (i.e., a respiratory ventilator, a patient monitor for intensive care, and an endoscope). The overall results showed (i) an accuracy of the data communication greater than 99 %, along with (ii) an average time response below ms, and (iii) satisfying feedback from the SUS questionnaires filled out by the physicians after intensive use.

This work addresses an innovative processing strategy to improve the classification of Steady-State Visually Evoked Potentials (SSVEPs). This strategy resorts to the combined use of fast Fourier transform and Canonical Correlation Analysis in time domain, and manages to outperform by over 5% previous results obtained for highly wearable, single-channel Brain–Computer Interfaces. In fact, a classification accuracy of 90% is reached with only 2-s time response. Then, the proposed algorithm is employed for an experimental characterization of three different Augmented Reality (AR) devices (namely, Microsoft Hololens I, Epson Moverio BT-350, and Oculus Rift S). These devices are used to generate the flickering stimuli necessary to the SSVEP induction. Also, in the three pieces of instrumentation under test, the number of simultaneous visual stimuli was increased with respect to the state-of-art solutions. The aim of the experimental characterization was to evaluate the influence of different AR technologies on the elicitation of user's SSVEPs. Classification accuracy, time response, and information transfer rate were used as figures of merit on nine volunteers for each piece of instrumentation. Experimental results show that choosing an adequate AR headset is crucial for obtaining satisfying performance: in fact, it can be observed that the classification accuracy obtained with Microsoft Hololens is about 20% greater than Epson Moverio one.

An integrated real-time monitoring system based on Augmented Reality (AR) and Brain–Computer Interface (BCI) for hands-free acquisition and visualization of remote data is proposed. As a case study, the monitoring of patients’ vitals in the operating room (OR) is considered; in particular, through the suitable combination of BCI and AR, the anesthetist can monitor in real-time (through a set of AR glasses), the patient’s vitals acquired from the electromedical equipment. Healthcare-related applications are particularly demanding in terms of real-time requirements; hence, the considered scenario represents an interesting and challenging testbed for the proposed system. Experimental tests were carried out at the University Hospital Federico II (Naples, Italy), employing pieces of equipment that are generally available in the OR. After the preliminary functional validation, accuracy and delay were measured, demonstrating the effectiveness and reliability of the proposed AR-BCI-based monitoring system.

In this work, an augmented-reality (AR) system for monitoring patient’s vitals in real time during surgical procedures is proposed and metrologically characterized in terms of transmission error rates and latency. These specifications, in fact, are crucial to ensure real-time response which is a critical requirement for Health applications. The proposed system automatically acquires the vitals from the operating room instrumentation, and shows them in real time directly on a set of wearable AR glasses. In this way, the surgical team has a real-time visualization of a comprehensive set of patient’s information, without constantly looking at the instrumentation. The system was designed to ensure modularity, flexibility, ease of use and, most importantly, a reliable communication. The system was designed, implemented and validated experimentally through on-field tests carried out at the Academic Hospital of Federico II University, using instrumentation typically available in the operating room: namely, a respiratory ventilator and a patient monitor. Results of the experimental validation highlight the need for the metrological approach in conceiving advanced monitoring surgical procedures in operating room.

In this work, an innovative system architecture for enhancing patient’s health monitoring during surgical procedure is proposed. The proposed integrated system consists of a video see-through (VST) headset (worn by anesthetist or nurses or other member of the surgical team involved in the procedure), which shows in real-time a comprehensive set of information on the patient’s health status.In particular, the operator can visualize in real-time the patient’s vitals acquired from the operating room (OR) equipment and access the electronic medical records. The proposed system was designed in view of robotic surgery; therefore, an additional function is that the VST will display the real-time view of what the surgeon is looking at on his console. Finally, the system will also include a machine-learning algorithm which, starting from the medical records and from the monitored vitals, will predict possible aggravating conditions of the patient’s health and provide a timely alert.The present work focuses on creating the implementation framework of the whole system, and on real-time monitoring of the patient’s vitals.

This work addresses the design, development and implementation of a 4.0-based wearable soft transducer for patient-centered vitals telemonitoring. In particular, first, the soft transducer measures hypertension-related vitals (heart rate, oxygen saturation and systolic/diastolic pressure) and sends the data to a remote database (which can be easily consulted both by the patient and the physician). In addition to this, a dedicated deep learning algorithm, based on a Long-Short-Term-Memory Autoencoder, was designed, implemented and tested for providing an alert when the patient’s vitals exceed certain thresholds, which are automatically personalized for the specific patient. Furthermore, a mobile application (EcO2u) was developed to manage the entire data flow and facilitate the data fruition; this application also implements an innovative face-detection algorithm that ensures the identity of the patient. The robustness of the proposed soft transducer was validated experimentally on five individuals, who used the system for 30 days. The experimental results demonstrated an accuracy in anomaly detection greater than 93%, with a true positive rate of more than 94%.

A highly-wearable single-channel Brain- Computer Interface (BCI) based on Steady-State Visually Evoked Potentials (SSVEPs) and Augmented Really (AR) is proposed. The SSVEP elicitation is provided by three AR head-mounted displays (HMD), namely Epson Moverio BT-350, Oculus Rift S, and Microsoft HoloLens. Four flickering stimuli, ranging from 8 Hz to 15 Hz, are used. The goal of the work is to carry out a performance comparison of the three aforementioned devices, in terms of stimuli visualization and SSVEPs detection. To this aim, classification accuracy and time response were assessed involving nine healthy volunteers during the experimental activity. The obtained results demonstrate that choosing an adequate HMD to render the flickering stimuli is decisive for obtaining adequate performances.