Peer reviewed journal papers

G. Valenza, L. Citi, E. P. Scilingo, and R. Barbieri, “Point-process nonlinear models with Laguerre and Volterra expansions: Instantaneous assessment of heartbeat dynamics,” IEEE Transactions on Signal Processing, p. (in press), 2013. [ bib | publisher ]

In the last decades, mathematical modelling and signal processing techniques have played an important role in the study of cardiovascular control physiology and heartbeat nonlinear dynamics. In particular, nonlinear models have been devised for the assessment of the cardiovascular system by accounting for short-memory second-order nonlinearities. In this paper, we introduce a novel inverse Gaussian point process model with Laguerre expansion of the nonlinear Volterra kernels. Within the model, the second-order nonlinearities also account for the long-term information given by the past events of the nonstationary non-Gaussian time series. In addition, the mathematical link to an equivalent cubic input-output Wiener-Volterra model allows for a novel instantaneous estimation of the dynamic spectrum, bispectrum and trispectrum of the considered inter-event intervals. The proposed framework is tested with synthetic simulations and two experimental heartbeat interval datasets. Applications on further heterogeneous datasets such as milling inserts, neural spikes, gait from short walks, and geyser geologic events are also reported. Results show that our model improves on previously developed models and, at the same time, it is able to provide a novel instantaneous characterization and tracking of the inherent nonlinearity of heartbeat dynamics.

L. Citi, E. N. Brown, and R. Barbieri, “A real-time automated point-process method for the detection and correction of erroneous and ectopic heartbeats.,” IEEE Transactions on Biomedical Engineering, vol. 59, pp. 2828-2837, Oct. 2012. [ bib | publisher ]

The presence of recurring arrhythmic events (also known as cardiac dysrhythmia or irregular heartbeats), as well as erroneous beat detection due to low signal quality, significantly affects estimation of both time and frequency domain indices of heart rate variability (HRV). A reliable, real-time classification and correction of ECG-derived heartbeats is a necessary prerequisite for an accurate online monitoring of HRV and cardiovascular control. We have developed a novel point-process-based method for real-time R-R interval error detection and correction. Given an R-wave event, we assume that the length of the next R-R interval follows a physiologically motivated, time-varying inverse Gaussian probability distribution. We then devise an instantaneous automated detection and correction procedure for erroneous and arrhythmic beats by using the information on the probability of occurrence of the observed beat provided by the model. We test our algorithm over two datasets from the PhysioNet archive. The Fantasia normal rhythm database is artificially corrupted with known erroneous beats to test both the detection procedure and correction procedure. The benchmark MIT-BIH Arrhythmia database is further considered to test the detection procedure of real arrhythmic events and compare it with results from previously published algorithms. Our automated algorithm represents an improvement over previous procedures, with best specificity for the detection of correct beats, as well as highest sensitivity to missed and extra beats, artificially misplaced beats, and for real arrhythmic events. A near-optimal heartbeat classification and correction, together with the ability to adapt to time-varying changes of heartbeat dynamics in an online fashion, may provide a solid base for building a more reliable real-time HRV monitoring device.

M. Tombini, J. Rigosa, F. Zappasodi, C. Porcaro, L. Citi, J. Carpaneto, P. M. Rossini, and S. Micera, “Combined analysis of cortical (EEG) and nerve stump signals improves robotic hand control.,” Neurorehabil Neural Repair, vol. 26, pp. 275-281, July 2012. [ bib | publisher ]

BACKGROUND: Interfacing an amputee's upper-extremity stump nerves to control a robotic hand requires training of the individual and algorithms to process interactions between cortical and peripheral signals. OBJECTIVE: To evaluate for the first time whether EEG-driven analysis of peripheral neural signals as an amputee practices could improve the classification of motor commands. METHODS: Four thin-film longitudinal intrafascicular electrodes (tf-LIFEs-4) were implanted in the median and ulnar nerves of the stump in the distal upper arm for 4 weeks. Artificial intelligence classifiers were implemented to analyze LIFE signals recorded while the participant tried to perform 3 different hand and finger movements as pictures representing these tasks were randomly presented on a screen. In the final week, the participant was trained to perform the same movements with a robotic hand prosthesis through modulation of tf-LIFE-4 signals. To improve the classification performance, an event-related desynchronization/synchronization (ERD/ERS) procedure was applied to EEG data to identify the exact timing of each motor command. RESULTS: Real-time control of neural (motor) output was achieved by the participant. By focusing electroneurographic (ENG) signal analysis in an EEG-driven time window, movement classification performance improved. After training, the participant regained normal modulation of background rhythms for movement preparation (α/β band desynchronization) in the sensorimotor area contralateral to the missing limb. Moreover, coherence analysis found a restored α band synchronization of Rolandic area with frontal and parietal ipsilateral regions, similar to that observed in the opposite hemisphere for movement of the intact hand. Of note, phantom limb pain (PLP) resolved for several months. CONCLUSIONS: Combining information from both cortical (EEG) and stump nerve (ENG) signals improved the classification performance compared with tf-LIFE signals processing alone; training led to cortical reorganization and mitigation of PLP.

C. Balocco, V. Gori, E. Marmonti, and L. Citi, “Building-plant system energy sustainability. an approach for transient thermal performance analysis,” Energy and Buildings, vol. 49, pp. 443-453, June 2012. [ bib | publisher ]

The aim of this research is to investigate the thermophysical behaviour of building envelope multi-layer components and assess the whole building-plant response. The investigation was carried out in two different phases. The first phase concerned the study of the dynamic response of building envelope multi-layer components, by providing an improved implementation of the present European Standard UNI EN ISO 13786:2008. The proposed method allows one to determine the transient thermophysical performance of opaque building components that are associated with a thermal zone maintained at constant or variable air temperature. The variable internal air temperature can be associated with experimental data. The second phase consisted of a transient analysis of the above thermal zone with a conditioning plant working. The building plant simulations were carried out by EnergyPlus software. The method proposed, based on European standards, is particularly useful in the perspective of the Certification of the Energy Performance of the building-plant system, required by the European Union. It can provide a useful tool for developing an abacus of different envelope multi-layer component performances in order to support design choices.

M. Salvaris, C. Cinel, L. Citi, and R. Poli, “Novel protocols for P300-based brain-computer interfaces,” IEEE Transactions on Neural System and Rehabilitation Engineering, vol. 20, pp. 8-17, Jan. 2012. [ bib | publisher | .pdf ]

The oddball protocol is often used in Brain- Computer Interfaces (BCIs) to induce P300 ERPs, although, recently, some issues have been shown to detrimentally effect its performance. In this paper, we study a new periodic protocol and explore whether it can compete with the standard oddball protocol within the context of a BCI mouse. We found that the new protocol consistently and significantly outperforms the standard oddball protocol in relation to information transfer rates (33 bits/min for the former and 22 bits/min for the latter, measured at 90 we performed a comparison of two periodic protocols with two less conventional oddball-like protocols that reveals the importance of the interactions between task and sequence in determining the success of a protocol.

S. Micera, P. M. Rossini, J. Rigosa, L. Citi, J. Carpaneto, S. Raspopovic, M. Tombini, C. Cipriani, G. Assenza, M. C. Carrozza, K.-P. Hoffmann, K. Yoshida, X. Navarro, and P. Dario, “Decoding of grasping information from neural signals recorded using peripheral intrafascicular interfaces.,” Journal of Neuroengineering and Rehabilitation, vol. 8, p. 53, Sept. 2011. [ bib | publisher ]

BACKGROUND: The restoration of complex hand functions by creating a novel bidirectional link between the nervous system and a dexterous hand prosthesis is currently pursued by several groups. This connection must be fast, intuitive, and quite natural to allow an effective bidirectional flow of information between the user's nervous system and the smart artificial device. This goal can be achieved with several approaches and among them, the use of implantable interfaces with the peripheral nervous system, namely intrafascicular electrodes, is considered particularly interesting. METHODS: Thin-film longitudinal intra-fascicular electrodes were implanted in the median and ulnar nerves of an amputee's residuum during a four-week trial. The possibility of decoding motor commands suitable to control a dexterous hand prosthesis was investigated for the first time in this research field by implementing a spike sorting and classification algorithm. RESULTS: The results showed that motor information (e.g., grip types and single finger movements) could be extracted with classification accuracy around 85 (for three classes plus rest) and that the user could improve his ability to govern motor commands over time as shown by the improved discrimination ability of our classification algorithm. CONCLUSIONS: These results open up new and promising possibilities for the development of a neuro-controlled hand prosthesis.

L. Citi, R. Poli, and C. Cinel, “Documenting, modelling and exploiting P300 amplitude changes due to variable target delays in Donchin's speller.,” Journal of Neural Engineering, vol. 7, p. 056006, Oct. 2010. [ bib | publisher | .pdf ]

The P300 is an endogenous event-related potential (ERP) that is naturally elicited by rare and significant external stimuli. P300s are used increasingly frequently in brain-computer interfaces (BCIs) because the users of ERP-based BCIs need no special training. However, P300 waves are hard to detect and, therefore, multiple target stimulus presentations are needed before an interface can make a reliable decision. While significant improvements have been made in the detection of P300s, no particular attention has been paid to the variability in shape and timing of P300 waves in BCIs. In this paper we start filling this gap by documenting, modelling and exploiting a modulation in the amplitude of P300s related to the number of non-targets preceding a target in a Donchin speller. The basic idea in our approach is to use an appropriately weighted average of the responses produced by a classifier during multiple stimulus presentations, instead of the traditional plain average. This makes it possible to weigh more heavily events that are likely to be more informative, thereby increasing the accuracy of classification. The optimal weights are determined through a mathematical model that precisely estimates the accuracy of our speller as well as the expected performance improvement w.r.t. the traditional approach. Tests with two independent datasets show that our approach provides a marked statistically significant improvement in accuracy over the top-performing algorithm presented in the literature to date. The method and the theoretical models we propose are general and can easily be used in other P300-based BCIs with minimal changes.

R. Poli, C. Cinel, L. Citi, and F. Sepulveda, “Reaction-time binning: a simple method for increasing the resolving power of erp averages.,” Psychophysiology, vol. 47, pp. 467-485, May 2010. [ bib | publisher | .pdf ]

Stimulus-locked, response-locked, and ERP-locked averaging are effective methods for reducing artifacts in ERP analysis. However, they suffer from a magnifying-glass effect: they increase the resolution of specific ERPs at the cost of blurring other ERPs. Here we propose an extremely simple technique-binning trials based on response times and then averaging-which can significantly alleviate the problems of other averaging methods. We have empirically evaluated the technique in an experiment where the task requires detecting a target in the presence of distractors. We have also studied the signal-to-noise ratio and the resolving power of averages with and without binning. Results indicate that the method produces clearer representations of ERPs than either stimulus-locked and response-locked averaging, revealing finer details of ERPs and helping in the evaluation of the amplitude and latency of ERP waves. The method is applicable to within-subject and between-subject averages.

P. M. Rossini, S. Micera, A. Benvenuto, J. Carpaneto, G. Cavallo, L. Citi, C. Cipriani, L. Denaro, V. Denaro, G. D. Pino, F. Ferreri, E. Guglielmelli, K.-P. Hoffmann, S. Raspopovic, J. Rigosa, L. Rossini, M. Tombini, and P. Dario, “Double nerve intraneural interface implant on a human amputee for robotic hand control.,” Clinical Neurophysiology, vol. 121, pp. 777-783, May 2010. [ bib | publisher | .pdf ]

OBJECTIVES: The principle underlying this project is that, despite nervous reorganization following upper limb amputation, original pathways and CNS relays partially maintain their function and can be exploited for interfacing prostheses. Aim of this study is to evaluate a novel peripheral intraneural multielectrode for multi-movement prosthesis control and for sensory feed-back, while assessing cortical reorganization following the re-acquired stream of data. METHODS: Four intrafascicular longitudinal flexible multielectrodes (tf-LIFE4) were implanted in the median and ulnar nerves of an amputee; they reliably recorded output signals for 4 weeks. Artificial intelligence classifiers were used off-line to analyse LIFE signals recorded during three distinct hand movements under voluntary order. RESULTS: Real-time control of motor output was achieved for the three actions. When applied off-line artificial intelligence reached >85% real-time correct classification of trials. Moreover, different types of current stimulation were determined to allow reproducible and localized hand/fingers sensations. Cortical organization was observed via TMS in parallel with partial resolution of symptoms due to the phantom-limb syndrome (PLS). CONCLUSIONS: tf-LIFE4s recorded output signals in human nerves for 4 weeks, though the efficacy of sensory stimulation decayed after 10 days. Recording from a number of fibres permitted a high percentage of distinct actions to be classified correctly. Reversal of plastic changes and alleviation of PLS represent corollary findings of potential therapeutic benefit. SIGNIFICANCE: This study represents a breakthrough in robotic hand use in amputees.

S. Micera, L. Citi, J. Rigosa, J. Carpaneto, S. Raspopovic, G. Di Pino, L. Rossini, K. Yoshida, L. Denaro, P. Dario, and P. Rossini, “Decoding information from neural signals recorded using intraneural electrodes: Toward the development of a neurocontrolled hand prosthesis,” Proceedings of the IEEE, vol. 98, pp. 407 -417, Mar. 2010. [ bib | publisher | .pdf ]

The possibility of controlling dexterous hand prostheses by using a direct connection with the nervous system is particularly interesting for the significant improvement of the quality of life of patients, which can derive from this achievement. Among the various approaches, peripheral nerve based intrafascicular electrodes are excellent neural interface candidates, representing an excellent compromise between high selectivity and relatively low invasiveness. Moreover, this approach has undergone preliminary testing in human volunteers and has shown promise. In this paper, we investigate whether the use of intrafascicular electrodes can be used to decode multiple sensory and motor information channels with the aim to develop a finite state algorithm that may be employed to control neuroprostheses and neurocontrolled hand prostheses. The results achieved both in animal and human experiments show that the combination of multiple sites recordings and advanced signal processing techniques (such as wavelet denoising and spike sorting algorithms) can be used to identify both sensory stimuli (in animal models) and motor commands (in a human volunteer). These findings have interesting implications, which should be investigated in future experiments.

R. Poli, L. Citi, F. Sepulveda, and C. Cinel, “Analogue evolutionary brain computer interfaces,” IEEE Computational Intelligence Magazine, vol. 4, pp. 27 -31, Nov. 2009. [ bib | publisher ]

The keyboard is a device that, provides an interface that is reliable but also very unnatural. The mouse is only slightly less primitive, being an electro-mechanical transducer of musculoskeletal movement. Both have been with us for decades, yet they are unusable for people with severe musculoskeletal disorders and are themselves known causes of work-related upper-limb and back disorders, both hugely widespread problems. It will be a major contribution to computer interface technology to replace mouse and keyboard with brain-computer interfaces (BCIs) capable of directly interpreting the desires and intentions of computer users. In this article we describe the approach, results and promising new research directions in the realization of BCIs, with particular reference to a 2D pointing device. Three features characterize the approach. Firstly, BCI is logically analogue, second is the use of evolutionary algorithms, and the third feature is its interdisciplinarity.

L. Citi, O. Tonet, and M. Marinelli, “Matching Brain-Machine Interface performance to space applications,” International Review of Neurobiology, vol. 86, pp. 199-212, July 2009. [ bib | publisher | .pdf ]

A brain-machine interface (BMI) is a particular class of human-machine interface (HMI). BMIs have so far been studied mostly as a communication means for people who have little or no voluntary control of muscle activity. For able-bodied users, such as astronauts, a BMI would only be practical if conceived as an augmenting interface. A method is presented for pointing out effective combinations of HMIs and applications of robotics and automation to space. Latency and throughput are selected as performance measures for a hybrid bionic system (HBS), that is, the combination of a user, a device, and a HMI. We classify and briefly describe HMIs and space applications and then compare the performance of classes of interfaces with the requirements of classes of applications, both in terms of latency and throughput. Regions of overlap correspond to effective combinations. Devices requiring simpler control, such as a rover, a robotic camera, or environmental controls are suitable to be driven by means of BMI technology. Free flyers and other devices with six degrees of freedom can be controlled, but only at low-interactivity levels. More demanding applications require conventional interfaces, although they could be controlled by BMIs once the same levels of performance as currently recorded in animal experiments are attained. Robotic arms and manipulators could be the next frontier for noninvasive BMIs. Integrating smart controllers in HBSs could improve interactivity and boost the use of BMI technology in space applications.

C. Menon, C. de Negueruela, J. del R. Millán, O. Tonet, F. Carpi, M. Broschart, P. Ferrez, A. Buttfield, F. Tecchio, F. Sepulveda, L. Citi, C. Laschi, M. Tombini, P. Dario, P. M. Rossini, and D. D. Rossi, “Prospects of brain-machine interfaces for space system control,” Acta Astronautica, vol. 64, pp. 448-456, Feb. 2009. [ bib | publisher | .pdf ]

The dream of controlling and guiding computer-based systems using human brain signals has slowly but steadily become a reality. The available technology allows real-time implementation of systems that measure neuronal activity, convert their signals, and translate their output for the purpose of controlling mechanical and electronic systems. This paper describes the state of the art of non-invasive brain-machine interfaces (BMIs) and critically investigates both the current technological limits and the future potential that BMIs have for space applications. We present an assessment of the advantages that BMIs can provide and justify the preferred candidate concepts for space applications together with a vision of future directions for their implementation.

R. Poli, N. F. McPhee, L. Citi, and E. Crane, “Memory with memory in genetic programming,” Journal of Artificial Evolution and Applications, vol. 2009, pp. 1-16, Jan. 2009. [ bib | publisher ]

We introduce Memory with Memory Genetic Programming (MwM-GP), where we use soft assignments and soft return operations. Instead of having the new value completely overwrite the old value of registers or memory, soft assignments combine such values. Similarly, in soft return operations the value of a function node is a blend between the result of a calculation and previously returned results. In extensive empirical tests, MwM-GP almost always does as well as traditional GP, while significantly outperforming it in several cases. MwM-GP also tends to be far more consistent than traditional GP. The data suggest that MwM-GP works by successively refining an approximate solution to the target problem and that it is much less likely to have truly ineffective code. MwM-GP can continue to improve over time, but it is less likely to get the sort of exact solution that one might find with traditional GP.

S. Micera, X. Navarro, J. Carpaneto, L. Citi, O. Tonet, P. M. Rossini, M. C. Carrozza, K.-P. Hoffmann, M. Vivó, K. Yoshida, and P. Dario, “On the use of longitudinal intrafascicular peripheral interfaces for the control of cybernetic hand prostheses in amputees.,” IEEE Transactions on Neural System and Rehabilitation Engineering, vol. 16, pp. 453-472, Oct. 2008. [ bib | publisher ]

Significant strides have been recently made to develop highly sensorized cybernetic prostheses aimed at restoring sensorimotor limb functions to those who have lost them because of a traumatic event (amputation). In these cases, one of the main goals is to create a bidirectional link between the artificial devices (e.g., robotic hands, arms, or legs) and the nervous system. Several human-machine interfaces (HMIs) are currently used to this aim. Among them, interfaces with the peripheral nervous system and in particular longitudinal intrafascicular electrodes can be a promising solution able to improve the current situation. In this paper, the potentials and limits of the use of this interface to control robotic devices are presented. Specific information is provided on: 1) the neurophysiological bases for the use peripheral nerve interfaces; 2) a comparison of the potentials of the different peripheral neural interfaces; 3) the possibility of extracting and appropriately interpreting the neural code for motor commands and of delivering sensory feedback by stimulating afferent fibers by using longitudinal intrafascicular electrodes; 4) a preliminary comparative analysis of the performance of this approach with the ones of others HMIs; 5) the open issues which have to be addressed for a chronic usability of this approach.

L. Citi, J. Carpaneto, K. Yoshida, K.-P. Hoffmann, K. P. Koch, P. Dario, and S. Micera, “On the use of wavelet denoising and spike sorting techniques to process electroneurographic signals recorded using intraneural electrodes.,” Journal of Neuroscience Methods, vol. 172, pp. 294-302, July 2008. [ bib | publisher | .pdf ]

Among the possible interfaces with the peripheral nervous system (PNS), intraneural electrodes represent an interesting solution for their potential advantages such as the possibility of extracting spikes from electroneurographic (ENG) signals. Their use could increase the precision and the amount of information which can be detected with respect to other processing methods. In this study, in order to verify this assumption, thin-film longitudinal intrafascicular electrodes (tfLIFE) were implanted in the sciatic nerve of rabbits. Various sensory stimuli were applied to the hind limb of the animal and the elicited ENG signals were recorded using the tfLIFEs. These signals were processed to determine whether the different types of information can be decoded. Signals were wavelet denoised and spike sorted. Support vector machines were trained to use the spike waveforms found to infer the stimulus applied to the rabbit. This approach was also compared with previously used ENG-processing methods. The results indicate that the combination of wavelet denoising and spike sorting techniques can increase the amount of information extractable from ENG signals recorded with intraneural electrodes. This strategy could allow the development of more effective closed-loop neuroprostheses and hybrid bionic systems connecting the human nervous system with artificial devices.

L. Citi, R. Poli, F. Sepulveda, and C. Cinel, “P300-based BCI mouse with genetically-optimized analogue control,” IEEE Transactions on Neural System and Rehabilitation Engineering, vol. 16, pp. 51-61, Feb. 2008. [ bib | publisher | .pdf ]

In this paper we propose a brain-computer interface (BCI) mouse based on P300 waves in electroencephalogram (EEG) signals. The system is analogue in that at no point a binary decision is made as to whether or not a P300 was actually produced in response to the stimuli. Instead, the 2-D motion of the pointer on the screen, using a novel BCI paradigm, is controlled by directly combining the amplitudes of the output produced by a filter in the presence of different stimuli. This filter and the features to be combined within it are optimised by an evolutionary algorithm.

O. Tonet, M. Marinelli, L. Citi, P. M. Rossini, L. Rossini, G. Megali, and P. Dario, “Defining Brain-Machine Interface applications by matching interface performance with device requirements.,” Journal of Neuroscience Methods, vol. 167, pp. 91-104, Jan. 2008. [ bib | publisher | .pdf ]

Interaction with machines is mediated by human-machine interfaces (HMIs). Brain-machine interfaces (BMIs) are a particular class of HMIs and have so far been studied as a communication means for people who have little or no voluntary control of muscle activity. In this context, low-performing interfaces can be considered as prosthetic applications. On the other hand, for able-bodied users, a BMI would only be practical if conceived as an augmenting interface. In this paper, a method is introduced for pointing out effective combinations of interfaces and devices for creating real-world applications. First, devices for domotics, rehabilitation and assistive robotics, and their requirements, in terms of throughput and latency, are described. Second, HMIs are classified and their performance described, still in terms of throughput and latency. Then device requirements are matched with performance of available interfaces. Simple rehabilitation and domotics devices can be easily controlled by means of BMI technology. Prosthetic hands and wheelchairs are suitable applications but do not attain optimal interactivity. Regarding humanoid robotics, the head and the trunk can be controlled by means of BMIs, while other parts require too much throughput. Robotic arms, which have been controlled by means of cortical invasive interfaces in animal studies, could be the next frontier for non-invasive BMIs. Combining smart controllers with BMIs could improve interactivity and boost BMI applications.

Book chapters

L. Citi and S. Micera, “Wavelet denoising and conditioning of neural recordings,” in Introduction to Neural Engineering for Motor Rehabilitation (D. Farina, W. Jensen, and M. Akay, eds.), ch. 9, IEEE/Wiley Press, 2013. [ bib ]

L. Citi, R. Poli, and C. Cinel, “High-significance averages of event-related potential via genetic programming,” in Genetic Programming Theory and Practice VII (R. Riolo, U.-M. O'Reilly, and T. McConaghy, eds.), Genetic and Evolutionary Computation, ch. 9, pp. 135-157, Springer US, 2010. [ bib | publisher ]

In this paper we use register-based genetic programming with memory-with memory to discover probabilistic membership functions that are used to divide up data-sets of event-related potentials recorded via EEG in psycho-physiological experiments based on the corresponding response times. The objective is to evolve membership functions which lead to maximising the statistical significance with which true brain waves can be reconstructed when averaging the trials in each bin. Results show that GP can significantly improve the fidelity with which ERP components can be recovered.

Peer reviewed conference proceedings

L. Citi and R. Barbieri, “PhysioNet 2012 challenge: Predicting mortality of ICU patients using a cascaded SVM-GLM paradigm,” in Computing in Cardiology, (Krakow), Sept. 2012. [ bib | .pdf ]

The focus of the PhysioNet/CinC Challenge 2012 is to develop methods for patient-specific prediction of in-hospital mortality using general descriptors recorded at the time of admission to the ICU and up to 37 time-series measurements collected during the first 48 hours after admission. We developed an algorithm that uses both general descriptors and time-series measurements to predict the in-hospital death (IHD) of ICU patients in Event 1, and to provide a probability estimate of IHD in Event 2. Both aggregated variables and general descriptors were used as features of quadratic Support Vector Machine (SVM) classifiers. Six SVMs were trained using, for each one, all the positive examples plus, in turn, one sixth of the negative examples in the training set. Finally, a Generalized Linear Model with probit link was used to predict the probability of IHD for Event 2 using the raw outputs of the six SVMs as regressors. A positive binary prediction of IHD for Event 1 was made when the probability estimate was higher than an optimized threshold. Official final results of the challenge reported that our entry achieved an Event 2 score of 17.8835, which is the best score out of the total 23 submissions, and Event 1 score of 0.534454 (second best score).

M. Orini, G. Valenza, L. Citi, and R. Barbieri, “Tetravariate point-process model for the continuous characterization of cardiovascular-respiratory dynamics during passive postural changes,” in Computing in Cardiology, (Krakow), Sept. 2012. [ bib | .pdf ]

In this study, we present a new methodology for time-varying characterization of cardiovascular (CV) control, which includes RR interval (RRI), systolic arterial pressure (SAP), respiration (RSP) and pulse transit time (PTT). Within a multivariate model, CV dynamics are represented as stochastic point processes whose means has a tetravariate autoregressive structure. Such framework provides the simultaneous time-frequency assessment of: (i) both arms of the SAP-RRI loop, along baroreflex and mechanical feedforward pathways; (ii) Respiratory sinus arrhythmia (RSA), through the direct evaluation of the interactions between RSP and the RRI; (iii) the coupling between cardiorespiratory activity and vascular tone through quantification of the interactions between PTT and the other CV variables. We validated the model by characterizing CV control in 16 healthy subjects during a tilt table test, and we were able to confirm a satisfactory model's goodness-of-fit. We further estimated transfer function gains, instantaneous powers and directed coherences, and observed that RSP strongly drove respiratory-related oscillations in all the other CV variables, and that PTT depended on RRI dynamics rather than blood pressure variations. During head-up tilt, baroreflex sensitivity and RSA decreased, while the gain from RRI to SAP increased, thus confirming previous physiological characterizations.

L. Citi, G. Valenza, and R. Barbieri, “Instantaneous estimation of high-order nonlinear heartbeat dynamics by Lyapunov exponents,” in Proceedings of the 34th IEEE Engineering in Medicine and Biology Society Conference, EMBC, (San Diego), pp. 13-16, 2012. [ bib | publisher ]

This paper introduces a novel methodology able to provide time varying estimates of the Lyapunov Spectrum within a point process framework. The algorithm is applied to ECG-derived data to characterize heartbeat nonlinear dynamics by using a cubic autoregressive point process model. Estimation of the model parameters is ensured by the Laguerre expansion of the Wiener-Volterra kernels along with a maximum local log-likelihood procedure. In addition to the instantaneous Lyapunov exponents, as well as indices related to higher order dynamic polyspectra, our method is also able to provide all the instantaneous time domain and frequency domain measures of instantaneous heart rate (HR) and heart rate variability (HRV) previously considered. Experimental results show that our method is able to track complex cardiovascular control dynamics during fast transitional gravitational changes.

L. Citi, G. Valenza, P. Purdon, E. Brown, and R. Barbieri, “Monitoring heartbeat nonlinear dynamics during general anesthesia by using the instantaneous dominant Lyapunov exponent,” in Proceedings of the 34th IEEE Engineering in Medicine and Biology Society Conference, EMBC, (San Diego), pp. 3124-3127, 2012. [ bib | publisher ]

We present a novel methodology for instantaneous estimation of quantitative correlates of depth of Anesthesia from noninvasive electrocardiographic recordings. The analysis is based on a point process model of heartbeat dynamics that allows for continuous tracking of linear and nonlinear HRV indices, including a novel instantaneous assessment of the Lyapunov Spectrum by using a cubic autoregressive formulation. The effective estimation of the model parameters is ensured by the Laguerre expansion of the Wiener-Volterra kernels along with the maximum local log-likelihood procedure. We apply the proposed assessment to experimental recordings from healthy subjects during propofol anesthesia. The new assessment reveals novel time-varying complex heartbeat dynamics that underlie the quasi-periodic heartbeat fluctuations elicited by the sympatho-vagal balance. Results suggest that such quantification provides important information which is independent from the standard autonomic assessment and significantly correlated with loss of consciousness. Further investigation will focus on evolving our mathematical approach towards a promising monitoring tool for an accurate, noninvasive assessment of general anesthesia.

M. Orini, L. Citi, and R. Barbieri, “Bivariate point process modeling and joint non-stationary analysis of pulse transit time and heart period,” in Proceedings of the 34th IEEE Engineering in Medicine and Biology Society Conference, EMBC, (San Diego), pp. 2831-2834, 2012. [ bib | publisher ]

Pulse transit time (PTT) is strictly related to pulse wave velocity and may be used for non-invasive monitoring of arterial stiffness and pressure, whose assessment is fundamental to detect cardiovascular dysfunctions. We propose a new model to characterize instantaneous PTT dynamics, and the interactions between PTT and R-R interval (RRI). In this model, PTT is described as a point process whose probability function depends on previous PTT and RRI values. From the model coefficients, instantaneous powers, coherence and directed coherence of each spectral component are estimated. We used this framework to study the changes that tilt table test provoked in PTT and RRI dynamics in 17 healthy subjects. Time-varying spectral and coherence analysis revealed that, although PTT and RRI were locally correlated, direct contribution of RRI on PTT was low during the entire test in high frequency band, and just after postural changes in low frequency band. We conclude that PTT may add valuable information for a more accurate characterization of cardiovascular regulation.

G. Valenza, L. Citi, E. Scilingo, and R. Barbieri, “Using Laguerre expansion within point-process models of heartbeat dynamics: A comparative study,” in Proceedings of the 34th IEEE Engineering in Medicine and Biology Society Conference, EMBC, (San Diego), pp. 29-32, 2012. [ bib | publisher ]

Point-process models have been recognized as a distinguished tool for the instantaneous assessment of heartbeat dynamics. Although not thoroughly linked to the physiology, nonlinear models also yield a more accurate quantification of cardiovascular control dynamics. Here, we propose a Laguerre expansion of the linear and nonlinear Wiener-Volterra kernels in order to account for the nonlinear and non-gaussian information contained in the ECG-derived heartbeat series while using a reduced number of parameters. Within an Inverse-Gaussian probability model, up to quadratic nonlinearities were considered to continuously estimate the dynamic spectrum and bispectrum. Results performed on 10 subjects undergoing a stand-up protocol show that this novel methodology improves on the algorithmic performances and, at the same time, more accurately characterizes sympatho-vagal changes to posture.

J. Carpaneto, A. Cutrone, S. Bossi, P. N. Sergi, L. Citi, J. Rigosa, P. M. Rossini, and S. Micera, “Activities on PNS neural interfaces for the control of hand prostheses,” in Proceedings of the 33rd IEEE Engineering in Medicine and Biology Society Conference, EMBC, (Boston), Sept. 2011. [ bib ]

The development of interfaces linking the human nervous system with artificial devices is an important area of research. Several groups are working on the development of devices able to restore sensory-motor function in subjects affected by neurological disorders, injuries or amputations. Neural electrodes implanted in peripheral nervous system, and in particular intrafascicular electrodes, seem to be a promising approach for the control of hand prosthesis thanks to the possibility to selectively access motor and sensory fibers for decoding motor commands and delivering sensory feedback. In this paper, activities on the use of PNS interfaces for the control of hand prosthesis are presented. In particular, the design and feasibility study of a self-opening neural interface is presented together with the decoding of ENG signals in one amputee to control a dexterous hand prosthesis.

Z. Chen, L. Citi, P. L. Purdon, E. N. Brown, and R. Barbieri, “Instantaneous assessment of autonomic cardiovascular control during general anesthesia,” in Proceedings of the 33rd IEEE Engineering in Medicine and Biology Society Conference, EMBC, (Boston), Sept. 2011. [ bib ]

We present a comprehensive probabilistic point process framework to estimate and monitor the instantaneous heartbeat dynamics as related to specific cardiovascular control mechanisms and hemodynamics. Assessment of the model's statistics is established through the Wiener-Volterra theory and a multivariate autoregressive (AR) structure. A variety of instantaneous cardiovascular metrics, such as heart rate (HR), heart rate variability (HRV), respiratory sinus arrhythmia (RSA), and baroreceptor-cardiac reflex (BRS), can be rigorously derived within a parametric framework and instantaneously updated with an adaptive algorithm. Instantaneous metrics of nonlinearity, such as the bispectrum of heartbeat intervals, can also be derived. We have applied the proposed point process framework to experimental recordings from healthy subjects in order to monitor cardiovascular regulation under propofol anesthesia. Results reveal interesting dynamic trends across different pharmacological interventions, confirming the ability of the algorithm to track important changes in cardiorespiratory elicited interactions, and pointing at our mathematical approach as a promising monitoring tool for an accurate, noninvasive assessment of general anesthesia.

L. Citi, M. T. Bianchi, E. B. Klerman, and R. Barbieri, “Instantaneous monitoring of sleep fragmentation by point process heart rate variability and respiratory dynamics,” in Proceedings of the 33rd IEEE Engineering in Medicine and Biology Society Conference, EMBC, (Boston), Sept. 2011. [ bib ]

We present a novel, automatic point-process approach that is able to provide continuous, instantaneous estimates of heart rate variability (HRV) and respiratory sinus arrhythmia (RSA) in long duration data recordings such as those during an entire night of sleep. We analyze subjects with and without sleep apnea who underwent diagnostic polysomnography. The proposed algorithm is able to quantify multi-scale high time resolution autonomic signatures of sleep fragmentation, such as arousals and stage transitions, throughout an entire night. Results demonstrate the ability of our methods to track fast dynamic transitions from sleep to wake and between REM sleep and other sleep stages, providing resolution details not available in sleep scoring summaries. An automatic threshold-based procedure is further able to detect brief arousals, with the instantaneous indices characterizing specific arousal dynamic signatures.

L. Citi, E. N. Brown, and R. Barbieri, “A point process local likelihood algorithm for robust and automated heart beat detection and correction,” in Proceedings of Computing in Cardiology, (Hangzhou), Sept. 2011. [ bib ]

Robust and automated classification and correction of ECG-derived heart beats are a necessary prerequisite for an accurate real-time estimation of measures of heart rate variability and cardiovascular control. In particular, the low quality of the signal, as well as the presence of recurring arrhythmic events, may significantly affect estimation accuracy. We here present a novel point process based method for a real time R-R interval error detection and correction. Results of detection analysis over data from the benchmark MIT-BIH arrhythmia database demonstrate that the proposed algorithm achieves 99.97% accuracy (98.23% sensitivity, 99.98% specificity and 95.69% positive predictive value), outperforming state-of-the-art algorithms. Further results on simulated data demonstrate the efficacy of the detection and correction method.

L. Citi, M. Djilas, C. Azevedo-Coste, K. Yoshida, E. N. Brown, and R. Barbieri, “Point-process analysis of neural spiking activity of muscle spindles recorded from thin-film longitudinal intrafascicular electrodes,” in Proceedings of the 33rd IEEE Engineering in Medicine and Biology Society Conference, EMBC, (Boston), pp. 2311-2314, Sept. 2011. [ bib | publisher | .pdf ]

Recordings from thin-film Longitudinal Intra-Fascicular Electrodes (tfLIFE) together with a wavelet-based denoising and a correlation-based spike sorting algorithm, give access to firing patterns of muscle spindle afferents. In this study we use a point process probability structure to assess mechanical stimulus-response characteristics of muscle spindle spike trains. We assume that the stimulus intensity is primarily a linear combination of the spontaneous firing rate, the muscle extension, and the stretch velocity. By using the ability of the point process framework to provide an objective goodness of fit analysis, we were able to distinguish two classes of spike clusters with different statistical structure. We found that spike clusters with higher SNR have a temporal structure that can be fitted by an inverse Gaussian distribution while lower SNR clusters follow a Poisson-like distribution. The point process algorithm is further able to provide the instantaneous intensity function associated with the stimulus-response model with the best goodness of fit. This important result is a first step towards a point process decoding algorithm to estimate the muscle length and possibly provide closed loop Functional Electrical Stimulation (FES) systems with natural sensory feedback information.

R. J. Ellis, L. Citi, and R. Barbieri, “A point process approach for analyzing gait variability dynamics,” in Proceedings of the 33rd IEEE Engineering in Medicine and Biology Society Conference, EMBC, (Boston), Sept. 2011. [ bib ]

We present a novel statistical paradigm for modeling and analysis of gait variability which captures the natural point process structure of gait intervals and allows for definition of new measures instantaneous mean and standard deviation. We validate our model using two existing data sets from physionet.org. Results show an excellent model fit and yield insights into the underlying statistical structure behind human gait. Statistical analyses further corroborate previous findings of increased variability in gait at different speeds, both self-paced and metronome-paced, and reveal a significant increase in gait variability in Parkinson's subjects, as compared to young and elderly healthy subjects. These results indicate the validity of a point process approach to the analysis of gait, and the potential utility of incorporating instantaneous measures of gait into diagnostic or patient monitoring applications.

D. Pani, F. Usai, L. Citi, and L. Raffo, “Real-time processing of tfLIFE neural signals on embedded DSP platforms: A case study,” in Proceedings of the 5th IEEE/EMBS International Conference on Neural Engineering, NER, (Cancun), pp. 44-47, Apr. 2011. [ bib | publisher ]

Spike sorting is a typical neural processing technique aimed at identifying the firing activity of individual neurons. It plays a different role in the processing of the signals coming either from a single electrode or an electrode array. In presence of highly noisy recordings, a preliminary denoising stage is required in order to improve the SNR. Despite the significant number of studies in the field, only a few of them deal with peripheral nervous system (PNS) recordings and often the possibility of a real-time implementation is only hinted without any real implementation study. In this paper, a real-time PNS signal processing and classification technique is presented end evaluated on real elec-troneurographic signals taken from the sciatic nerve of rats. A state-of-the-art algorithm, composed of a wavelet denoising preprocessing stage followed by a correlation-based spike sorting and a support vector machine, has been adapted to work on-line in order to improve the processing efficiency while preserving at the most its effectiveness. The algorithm provides some level of adaptiveness with respect to an off-line implementation. On average, the correct classification reach 92.24 can be easily filtered out. Cycle-accurate profiling results on an off-the-shelf Digital Signal Processor demonstrate the real-time performance.

R. Poli, C. Cinel, L. Citi, and M. Salvaris, “A genetic programming approach to detecting artifact-generating eye movements from eeg in the absence of electro-oculogram,” in Proceedings of the 5th IEEE/EMBS International Conference on Neural Engineering, NER, (Cancun), pp. 416-421, Apr. 2011. [ bib | publisher ]

In this paper we use genetic programming an evolutionary program-induction technology to evolve algorithms that accurately approximate the behaviour of two standard detectors of ocular movement based on Electro-oculogram (EOG). The prediction is based entirely on EEG signals, i.e., without using EOG, making it possible to detect eye movements even in data recorded without EOG or eye tracking. Experimental results with this approach are very encouraging.

L. Citi, E. B. Klerman, E. N. Brown, and R. Barbieri, “Point process heart rate variability assessment during sleep deprivation,” in Proceedings of Computing in Cardiology, (Belfast), Sept. 2010. [ bib | .pdf ]

To investigate the potential relationships between Heart rate variability (HRV) and objective performance-subjective alertness measures during sleep deprivation, a novel point process algorithm was applied to ECG data from healthy young subjects in a 52-hour Constant Routine protocol, which includes sleep deprivation. Our algorithm is able to estimate the time-varying behavior of the HRV spectral indexes in an on-line instantaneous fashion. Results demonstrate the ability of our framework to provide high time-resolution sympatho-vagal dynamics as measured by spectral low frequency (LF) and high frequency (HF) power. Correlation analysis on individual subjects reveals a relevant correspondence between LF/HF and subjective alertness during the initial hours of sleep deprivation. At longer times awake, high correlation levels between LF/HF and objective performance indicate an increasing sympathetic drive as performance measures worsen. These results suggest that our point-process based HRV assessment could aid in real-time prediction of performance-alertness.

R. Poli, L. Citi, M. Salvaris, C. Cinel, and F. Sepulveda, “Eigenbrains: the free vibrational modes of the brain as a new representation for EEG,” in Proceedings of the 32nd IEEE Engineering in Medicine and Biology Society Conference, EMBC, vol. 2010, (Buenos Aires), pp. 6011-6014, Sept. 2010. [ bib | publisher ]

We present a new transform for EEG signals whose basis functions are well suited to represent the large-scale dynamics associated with event related potentials. The method involves instantiating an approximate model of the electrical properties of the brain as a conductor medium and then studying the free vibrational modes of the model. These form a set of basis functions, which we call eigenbrains, that can be used to meaningfully re-represent the brain's electrical activity. Eigenbrains are compared to principal component analysis and independent component analysis to highlight differences and similarities.

M. Salvaris, C. Cinel, R. Poli, L. Citi, and F. Sepulveda, “Exploring multiple protocols for a brain-computer interface mouse,” in Proceedings of the 32nd IEEE Engineering in Medicine and Biology Society Conference, EMBC, (Buenos Aires), pp. 4189-4192, Sept. 2010. [ bib | publisher ]

In recent years, various visual protocols have been explored for P300-based BCI. In stimulus-driven BCI paradigms such as P300 BCIs it is vital to optimise the stimulation protocol as much as possible in order to achieve the best performance. Due to the inherent variability between subjects and the complex nature of the brain it is unlikely that an optimal protocol will be identified through a single iteration of random exploration. That is why in this paper we explore 8 different visual protocol configurations based on recent literature, in the hope of identifying key features that can later be used to create further improved protocols. Results indicate that luminosity changes, the standard method of stimulation used in visual P300 BCI protocols, do provide the best performance of the variations presented here.

S. Micera, J. Rigosa, J. Carpaneto, L. Citi, S. Raspopovic, E. Guglielmelli, A. Benvenuto, L. Rossini, G. D. Pino, G. Cavallo, M. C. Carrozza, C. Cipriani, K. P. Hoffmann, P. Dario, and P. M. Rossini, “On the control of a robot hand by extracting neural signals from the PNS: preliminary results from a human implantation.,” in Proceedings of the 31st IEEE Engineering in Medicine and Biology Society Conference, EMBC, vol. 2009, (Minneapolis), pp. 4586-4589, Sept. 2009. [ bib | publisher ]

The development of hybrid neuroprosthetic systems (HBSs) linking the human nervous system with artificial devices is an important area of research that is currently addressed by several groups to restore sensorimotor function in people affected by different disabilities. It is particularly important to establish a fast, intuitive, bidirectional flow of information between the nervous system of the user and the smart robotic device. Among the possible solutions to achieve this goal, interfaces with the peripheral nervous system and in particular intraneural electrodes can represent an interesting choice. In the present study, thin-film longitudinal intra-fascicular electrodes were implanted in the median and ulnar nerves of an amputee. The possibility of restoring the bidirectional link between the subject and the external world was investigated during a 4 week trial. The result showed that both the extraction of motor information and the restoration of sensory function are possible.

L. Citi, R. Poli, and C. Cinel, “Exploiting P300 amplitude variations can improve classification accuracy in Donchin's BCI speller,” in Proceedings of the 3rd IEEE/EMBS International Conference on Neural Engineering, NER, (Antalya), Apr. 2009. [ bib | publisher ]

The P300 is an endogenous component of EEG event related potentials which is elicited by rare and significant stimuli. P300s are used increasingly frequently in Brain Computer Interfaces (BCI) because, being naturally elicited in response to external stimuli, users do not need special training. However, P300 waves are hard to detect and, therefore, multiple stimulus presentations are needed before an interface can make a reliable decision. While significant improvements have been made in the detection of P300s, no particular attention has been paid to the variability in shape and timing of P300 waves and its exploitation in BCI. In this paper we start filling this gap, by first documenting and then exploiting a modulation in amplitude of P300 caused by target-to-target interval (TTI) differences. We demonstrate this within the context of the Donchin's speller, which is perhaps the best known example of a BCI system relying on the detection P300 waves, where target-to-target interval variations are induced by stimuli randomisation. In particular we show that by specialising detectors to work with P300s elicited with each TTI, we can further improve the performance of the best known Donchin's speller with minimal changes.

R. Poli, N. F. Mcphee, L. Citi, and E. Crane, “Memory with memory in tree-based genetic programming,” in Proceedings of the 12th European Conference on Genetic Programming, EuroGP '09, (Tübingen), pp. 25-36, Apr. 2009. [ bib | publisher ]

In recent work on linear register-based genetic programming (GP) we introduced the notion of Memory-with-Memory (MwM), where the results of operations are stored in registers using a form of soft assignment which blends a result into the current content of a register rather than entirely replace it. The MwM system yielded very promising results on a set of symbolic regression problems. In this paper, we propose a way of introducing MwM style behaviour in tree-based GP systems. The technique requires only very minor modifications to existing code, and, therefore, is easy to apply. Experiments on a variety of synthetic and real-world problems show that MwM is very beneficial in tree-based GP, too.

J. DiGiovanna, L. Citi, K. Yoshida, J. Carpaneto, J. C. Principe, J. C. Sanchez, and S. Micera, “Inferring the stability of life through brain machine interfaces.,” in Proceedings of the 30th IEEE Engineering in Medicine and Biology Society Conference, EMBC, vol. 2008, (Vancouver), pp. 2008-2011, Aug. 2008. [ bib | publisher ]

We examine neural signals from Longitudinally implanted Intra-Fascicular Electrodes (LIFE) in a chronic, rabbit model. Translation-invariant wavelet de-noising methods are used to improve S%R. Then traditional template-based spike sorting is applied to discriminate single units. We investigate the effect of discriminating between identified units on Brain Machine Interface (BMI) decoding performance. We infer the stability of LIFE based on decoding performance with and without current BMI methods to counter-act electrode neural signal degradation.

K. Yoshida, M. Kurstjens, L. Citi, K. Koch, and S. Micera, “Recording experience with the thin-film longitudinal intra-fascicular electrode, a multichannel peripheral nerve interface,” in Proceedings of the 10th IEEE International Conference on Rehabilitation Robotics ICORR, (Noordwijk), pp. 862-867, June 2007. [ bib | publisher ]

This paper presents our experience evaluating a multi-channel peripheral nerve interface, the thin-film longitudinal intra-fascicular electrode (tfLIFE). One application for the tfLIFE is their potential use as a means to detect independent channels of volitional commands from the amputee. The neural interface would be required to be sufficiently selective to detect the activity of single motor nerve fibres within the nerve stump. Experiments were conducted in the acute rabbit model evaluate the recording characteristics of the tfLIFE array. Multiunit activity was recorded in response to mechanical stimulation of peripheral mechanoreceptors. In some channels in all experiments, large single unit spikes were clearly visible. These data were then processed to determine whether an artificial discriminator could be trained to detect and track activity from the multi unit recordings. We also tested whether inclusion of multi channel information could be used to improve the performance of the discriminator. Our preliminary results indicate the inclusion of multiple channels significantly improves the performance.

R. Poli, C. Cinel, L. Citi, and F. Sepulveda, “Evolutionary brain computer interfaces.,” in Proceedings of the EvoWorkshops, (Valencia), pp. 301-310, Apr. 2007. [ bib ]

We propose a BCI mouse and speller based on the manipulation of P300 waves in EEG signals. The 2-D motion of the pointer on the screen is controlled by directly combining the amplitudes of the output produced by a filter in the presence of different stimuli. This filter and the features to be combined within it are optimised by a GA.

C. Menon, C. de Negueruela, J. del R. Millán, O. Tonet, F. Carpi, M. Broschart, P. Ferrez, A. Buttfield, P. Dario, L. Citi, C. Laschi, M. Tombini, F. Sepulveda, R. Poli, R. Palaniappan, F. Tecchio, P. M. Rossini, and D. De Rossi, “Prospects of Brain-Machine Interfaces for space system control,” in Proceedings of the 57th International Astronautical Congress, (Valencia), Oct. 2006. [ bib ]

The dream of controlling and guiding computer-based systems using human brain signals has slowly but steadily become a reality. The available technology allows real-time implementation of systems that measure neuronal activity, convert their signals, and translate their output for the purpose of controlling mechanical and electronic systems. This paper describes the state of the art of non-invasive brain-machine interfaces (BMIs) and critically investigates both the current technological limits and the future potential that BMIs have for space applications. We present an assessment of the advantages that BMIs can provide and justify the preferred candidate concepts for space applications together with a vision of future directions for their implementation.

L. Citi, R. Poli, and C. Cinel, “Analogue P300-based BCI pointing device,” in Proceedings of the 3rd International BCI Workshop and Training Course, (Graz), pp. 92-93, Sept. 2006. [ bib | .pdf ]

We propose a P300-based BCI mouse. The system is analogue: the pointer is controlled by directly combining the amplitudes of the outputs produced by a filter in the presence of different stimuli. The system is optimised by a genetic algorithm.

S. Micera, P. Sergi, J. Carpaneto, L. Citi, S. Bossi, K.-P. Koch, K.-P. Hoffmann, A. Menciassi, K. Yoshida, and P. Dario, “Experiments on the development and use of a new generation of intra-neural electrodes to control robotic devices,” in Proceedings of the 28th IEEE Engineering in Medicine and Biology Society Conference, EMBC, (New York City), pp. 2940 -2943, Sept. 2006. [ bib | publisher ]

The development of interfaces linking the human nervous system with artificial devices is an important area of research and several groups are now addressing it. Interfaces represent the key enabling technology for the development of devices usable for the restoration of motor and sensory function in subjects affected by neurological disorders, injuries or amputations. For example, current hand prostheses use electromyographic (EMG) signals to extract volitional commands but this limits the possibility of controlling several degrees of freedom and of delivering sensory feedback. To achieve these goals, implantable neural interfaces are required. Among the candidate interfaces with the peripheral nervous system intra-neural electrodes seem to be an interesting solution due to their bandwidth and ability to access volition and deliver sensory feedback. However, several drawbacks have to be addressed in order to increase their usability. In this paper, experiments to address many of these issues are presented as part of the development of a new generation of intra-neural electrodes. The results showed seem to confirm that these new interfaces seem to have interesting properties and that they can represent a significant improvement of the state of the art. Extensive experiments will be carried out in the future to validate these results

L. Citi, J. Carpaneto, K. Yoshida, K.-P. Hoffmann, K.-P. Koch, P. Dario, and S. Micera, “Characterization of tfLIFE neural response for the control of a cybernetic hand,” in Proceedings of the 1st IEEE/RAS-EMBS International Conference on Biomedical Robotics and Biomechatronics, BioRob, (Pisa), pp. 477 -482, Feb. 2006. [ bib | publisher ]

The development of interfaces that link the human nervous system with robotic devices or man-made devices has been a main area of research for several groups in the world. These groups focus on the restoration of motor and sensory function to those with degenerative diseases, injury or in amputees. A key component is these systems is a fast, intuitive, bidirectional interface between the biological and mechatronic systems that allows the robotic limb to be controlled as if it were a natural part of the body. Current hand prostheses use electromyographic (EMG) signals, but are limited to a small number of channels and to sensing volition. To achieve sensory feedback and a higher number of control channels, a neuroprosthetic interface are required. In the present study, thin-film longitudinal intra-fascicular electrodes (tfLIFE) were implanted in the sciatic nerve of the rabbit. Various sensory stimuli were applied to the hind limb of the rabbit and the elicited signals were recorded using the tfLIFEs. These signals were processed to determine whether the different modes of information could be decoded. Signals were Kalman filtered, wavelet denoised, and spike sorted. The classes of spikes found were then used to infer the stimulus applied to the rabbit. Although the signals acquired from a single tLIFE gave poor stimulus recognition, the combination of the signals from multiple sites led to better results. The spike sorting algorithm is also helped by the use of temporal correlation between the channels. A direct outcome of the results is the possibility of increasing the number of channels of control possible with a prosthetic limb

C. Cinel, R. Poli, and L. Citi, “Possible sources of perceptual errors in P300-based speller paradigm,” in Proceedings of the 2nd International BCI Workshop and Training Course, (Graz), pp. 39-40, Sept. 2004. [ bib | .pdf ]

Some perceptual phenomena can interfere with character identification in Farwell and Donchin's P300-based speller paradigm: attentional blink, repetition blindness and other effects caused by attentional limits. In the paper we discuss these and provide empirical evidence for one class of perceptual errors.

L. Citi, R. Poli, and F. Sepulveda, “An evolutionary approach to feature selection and classification in P300-based BCI,” in Proceedings of the 2nd International BCI workshop and Training Course, (Graz), pp. 41-42, Sept. 2004. [ bib | .pdf ]

We explore the use of evolutionary algorithms in the selection of features and the classification of P300 signals in BCI. As a result we have found new ways to process and combine EEG signals to improve detection.

Theses

L. Citi, Development of a neural interface for the control of a robotic hand. PhD thesis, IMT Lucca and Scuola Superiore Sant'Anna Pisa, Apr. 2009. [ bib | .pdf ]

The restoration of sensorimotor functions for the control ofartificial hands is a fundamental point in order to improve the quality of life of amputees. Current hand prostheses use electromyographic (EMG) signals, but are limited to a small number of channels and rely on visual feedback. As the interface is the bottleneck for a wide class of hybrid bionic systems, we have developed a general framework to match task requirements with interface performance. We have carefully and thoroughly examined the case of the amputee user and its requirements: throughput and latency, but also user-friendliness, invasiveness, bi-directionality, and possibility of natural control of the prosthesis. From this analysis, we have concluded that, albeit suboptimal from a mere throughput and latency point of view, peripheral invasive interface can represent a promising medium-term solution. We have compared the different types of interfaces with the peripheral nervous system, finding in longitudinal intra-fascicular interfaces (LIFEs) a tradeoff between invasiveness and selectivity. In order to assess the possibility of extracting complex information from LIFEs, we have run preliminary experiments with small animal models recording induced afferent information. Using the sophisticated signal processing techniques developed (wavelet denoising and spike sorting) and a robust classifier, we were able to discriminate four (or five) different classes of stimuli with performance in a range between 90% and 99%. These results confirmed and outperformed prior work carried out with different approaches. There are plans to validate the approach with a human amputee. Hence, several steps have been taken in order to make possible the recording of neural signals from a human subject and allow online processing and control of the “Cyberhand” smart robotic hand prosthesis.

L. Citi, “Un'interfaccia cervello-computer mediante metodi evoluzionistici di trattamento di segnali EEG,” Master's thesis, Università degli Studi di Firenze, July 2004. [ bib | .pdf ]

NOTICE: the [.pdf] link of some of the entries above allows the download of the author's version of a work that was accepted for publication in the corresponding journal. Changes resulting from the publishing process, such as peer review, editing, corrections, structural formatting, and other quality control mechanisms may not be reflected in the downloaded document. Changes may have been made to the work since it was submitted for publication. The definitive version published in the journal can be accessed by following the corresponding [publisher] link. In general, personal use of this material is permitted. Permission from publisher must be obtained for all other uses; for details, please follow the [publisher] link and seek further information on the publisher's website.