Specially-appointed adjunct associate professor
As a part of this appointment, in collaboration with Polish Academy of Science (IFJ-PAN), Kraków, Poland, work initiated in the previous year is being continued, with focus on the development of a hexapod robot controlled by central pattern generators and motor neurons simulated by means of non-linear circuits implement in Field-Programmable Analog Arrays (FPAAs).
Visiting professor, Principal specialist
He is a member of the research group of Prof. Stanisław Drożdż, and his work presently focuses on certain properties of bipolar junction-based electronic oscillators, namely searching for mono- and multi-fractal behaviour in the generated signals, and in general evaluating novel topologies for these oscillators and related properties. Continuing and extending previously-initiated research described below, further topics include i) mechanisms of remote synchronization and their study in terms of underlying dynamics, synchronization manifolds and information (perturbation) transfer, ii) emergent properties in complex networks, in particular self-similarity of links and paths (related to adjacency matrix idempotence) as a measure of system “collectivity”. As a part of this appointment, work was also done on brain-computer interfaces (in conjunction with Tokyo Institute of Technology, see below) and the lattice of glow lamps further described below. During this period, a body of work entitled “Electronic networks of chaotic oscillators as physical models of some emergent phenomena in neural dynamics” was completed, and submitted for consideration towards the title of Doktor Habilitowany in Physics.
Academic year 2016/2017, freelance/contract lecturer
Under this appointment, a second edition of the course entitled “Introduction to Robotics: an Experimental Approach” described below is being delivered, in a revised and expanded format.
In2H2 Inc., Folsom CA, USA
Non-disclosable contracted research and advising on large-scale data analysis applications, particularly in biomedicine and genetics, of vector distance calculation algorithms and in particular their hardware implementations at logic, board and system levels, involving collaboration with diverse entities and the filing of a patent.
Ongoing inventive activity
In addition to the other listed academic and professional activities, independent and self-funded applied research is being conducted around inventions in two areas: i) non-invasive, remote detection of liquid or solid excretion in sanitary towels by incontinent patients, based on a two-dimensional time-frequency impedance measurement implemented using a custom wearable wireless device of own design and a trivial, disposable textile element, ii) detection of anomalous events in household electric systems enabling substantially improved sensitivity to dangerous fault conditions compared to current circuit breakers, implemented by means of a custom-designed advanced multi-frequency measurement front-end. Two corresponding patents have been filed.
Aug 2015 - Jun 2016 - Guest post-doctoral fellow, funded as freelance via Scienze Mente-Cervello, Rovereto, Italy
Jul 2016 - current - Visiting researcher/guest fellow (honorary only)
Activities, now conducted in collaboration with Polish Academy of Science (IFJ-PAN), Kraków, Poland, focus on extending and confirming observations that scale-free network organization is associated with emergence of signatures of non-linear dynamics in the most strongly interconnected (“hub”) nodes. In particular, the results from previous work are being extended to the “meso-scale”, by means of analyzing electrophysiological data from in-vitro neural cultures on multi-electrode arrays, freely provided by the group of Prof. Steve Potter from Georgia Institute of Technology, USA. For this purpose, similar non-linear analyses to those detailed below are being utilized, however with recourse to specific techniques for generating surrogate spike-train time-series, and with tentative correlation to magnetoencephalography (MEG) data as well. Another area of work pertained to topological properties of complex networks, namely measures of self-similarity applicable across scales and diverse artificial and natural networks.
Work on the clinical resting-state fMRI portion of the PharmaCog project continued under supervision of Prof. Jorge Jovicich, alongside other projects on clinical applications of resting-state fMRI stemming from research previously conducted at the Fondazione Istituto Neurologico Carlo Besta.
Extending the work on non-linear circuits described below, a study was performed on a two-dimensional array of glow (neon) lamps, each one capacitively coupled to its neighbours. Avalanching behaviour having critical exponents 3/2 and 2 for size and duration respectively, as found in-vivo for entire brains and in-vitro for neural cultures, was observed, alongside hysteresis and metastability, pointing to a transition of the first kind.
Adjunct associate professor
In the course of this short-term visiting appointment, funded by the World Research Hub Initiative (WRHI) program of the Institute of Innovative Research, Tokyo Institute of Technology, he has worked in collaboration with and under the supervision of Prof. Yasuharu Koike and Prof. Natsue Yoshimura. The research topic was the development of brain-computer interface technologies (BCI), in particular consisting of the design and preliminary testing of a hybrid system combining a consumer-grade wearable headset acquiring an ensemble of bio-signals with a robotic manipulator guided by a stereoscopic vision system of own development. Experiments demonstrated successful control of the robot arm, and all data and source code of the completed project have been made public.
Academic year 2015/2016, freelance/contract lecturer
Under this appointment, he conceived and delivered a novel course entitled “Introduction to Robotics: an Experimental Approach”. This course, totalling 18 hours and 2 credits and delivered in the new “InfoBytes” format, was aimed at a multidisciplinary audience of students, and covered a broad range of topics including embedded computing and interfacing, analog and digital sensors, robot actuators and kinematics, machine vision principles and applications, and artificial intelligence. The course was innovative particularly in that all lessons revolved around practical experiments, utilizing a blend of diverse electronic platforms (particularly Arduino, MatLab and the Alan robotic platform of own design) to enhance engagement and learning.
Guest post-doctoral fellow
May 2013 - Jul 2015 | Fondazione Istituto Neurologico Carlo Besta, Milan, Italy
External collaborator, research fellow
Holding these posts jointly between the two institutions, research focused on applications of resting-state fMRI to the study of neurodegenerative disease, with two related technical strands being pursued. First, high-density connectivity mapping through replacing linear correlation with l1-norm in a massively parallel implementation via dedicated vector hardware of own design, on a PCI board based on the zero-instruction set computer (ZISC) architecture and released as open-source. Second, investigation of brain connectivity and non-linear dynamical properties to formulate novel, high-level bio-markers of network damage, in particular i) searching for non-linear, potentially chaotic, dynamics in “hub” regions based on clinical scans from the Alzheimer’s Disease Neuroimaging Initiative (ADNI) and the pan-European multi-centric PharmaCog project, ii) determining test-retest reliability and evaluating noise reduction strategies in resting-state fMRI connectivity, with activities conducted under the direction of Prof. Jorge Jovicich in the context of the PharmaCog project.
Work on brain signals intersected with independent, self-funded study of chaotic oscillators, aimed at developing a new research strand, with five related studies: 1) experimental characterization of the dynamics of five novel single-transistor oscillators, which despite their topological simplicity generate neural-like signals, 2) study of the emergence of phase- and amplitude-synchronization communities (modularity) on a ring of diffusively-coupled oscillators, 3) juxtaposing observations in the brain (based on data from Human Connectome Project) and an electronic network of chaotic oscillators, highlighting a similar relation between presence of hubs and dynamics in the two systems, 4) implementation of a related network as a CMOS circuit, suitable for implementing much larger networks, and generating chaos through quasi-periodicity by means of multiple ring oscillators of differing length, weakly coupled through diode/pass-gate cells (standard 0.7 μm process via IMEC-EuroPractice with support from Tecno77 Srl, Vicenza, Italy), 5) study of a dynamically-reconfigurable network implemented in the form of a dedicated PCI board, featuring 32 field-programmable analog arrays (FPAA, Anadigm Inc.) in addition to ADC, DAC and FPGA infrastructure, based on which spontaneous emergence of small-world topology via remote synchronization over a ring network was demonstrated. Particularly as they were conducted independently, the above activities provided the opportunity to gain deep insight into diverse aspects of analog and digital circuit design, programming and signal analysis, using a variety of software tools for SPICE simulation (LTSpice, T-Spice), PCB design (EdWin, PADS), CMOS Design (S-Edit, L-Edit), FPGA design (Xilinx ISE); for signal analysis the bulk of work was done in MatLab using a combination of own and repository code, with occasional recourse to C for interfacing and computationally-intensive algorithms.
Further work for the Fondazione Istituto Neurologico Carlo Besta involved analysing neuroimaging data from diverse patient populations and experimental models, in particular with reference to the anatomical and functional correlates of disorders of consciousness, pre-surgical applications of resting-state functional MRI, quantitative 1H-MR spectroscopy of 2-hydroxyglutarate and the study of demyelination in an experimental model of multiple sclerosis assessed using quantitative magnetization transfer-imaging at 7 Tesla.
In parallel with the academic activities, throughout this period he additionally conducted hardware development activities for a private start-up company, without financial remuneration and only for the purpose of professional development, namely for the possibility to experiment with leading-edge technologies in digital design and related issues. In this context, he independently designed a massively-parallel co-processor board based on the PCI express bus, including a high density Xilinx FPGA and an array of proprietary application-specific ICs (ASICs), choosing optimal solutions to the specific challenges raised by the application (particularly in terms of data bandwidth, clock, power and thermal management) and independently defining and designing the schematics of the board. He headed completion of the project, liaising with Tecno77 Srl (Vicenza, Italy) for all aspects of PCB layout and prototype manufacturing, further taking care of testing (via JTAG vectors as well as dedicated VHDL code) and characterization of prototypes. This work has also included advising on the interfacing standards and periphery design for the ASIC.
Research fellow, grade 8
He has been part of Dr. Dennis Chan’s research group, focusing on functional and structural neuroimaging of dementias and senile cognitive impairment. An area of key technical interest was the development of techniques for high-resolution mapping of resting-state functional connectivity using functional MRI.
In this context, he has designed a high-performance computational platform enabling high-resolution mapping of intrinsic connectivity, including a custom-designed co-processor board mounting SHARC ADSP-21369 (Analog Devices, Inc.) digital signal processors, which has been released as open source hardware and software. He has also developed an associated geodesic mapping platform, which involved a hardware implementation of Dijkstra’s algorithm using the VHDL language, deployed on a Virtex-7 (Xilinx, Inc.) field-programmable gate array (FPGA). Related numerical techniques have been applied in an exploratory study based on a dataset of patients with mild cognitive impairment scanned from Dr M. Bozzali’s group at the Fondazione IRCCS Santa Lucia (Rome, IT), demonstrating the superiority of whole-brain graph theoretical mapping to the use of source-separation techniques.
He has also conducted a study aiming to map the cognitive architecture underlying rapid calculation in an expert “human calculator”, has been involved in simultaneous PET/MRI addressing the co-localization and dissociation among structural, connectivity and metabolic changes in patients with Alzheimer’s disease and fronto-temporal dementia, and in clinical trials of a behavioural test of allocentric spatial memory to probe Alzheimer-related degeneration. He was a named researcher on a population scanning project funded by the UK government Technology Strategy Board (TSB) and Medical Research Council (MRC), and has been co-supervisor of two doctoral students.
Jun 2010 - Feb 2012 – Collaborator, full-time research fellow and group leader
Feb 2012 - Apr 2013 - External collaborator
In the context of his doctoral research with Prof. Critchley (see below), he focused on the study of decision-making under risk, investigating the representation of multiple economic parameters using simple prospects. Experiments included: i) exploring the univariate representation of. expected value, risk, uncertainty through fMRI, event-related potentials (ERPs), event-related spectral changes and autonomic responses, ii) behavioural study of patients with Parkinson’s disease and Huntington’s disease to evaluate the involvement of the basal ganglia in expected value representation, iii) exploring the relationship between prefrontal activity and risk taking via a neuromodulation experiment conducted using transcranial direct-current stimulation (tDCS) with the Universidate Presbiteriana MacKenzie (São Paulo, Brasile). He was also involved in other ERP studies with this laboratory, on social cognition and decision-making, semantic processing and autism, and action planning.
Other focus topics included i) development of algorithms and techniques to model functional connectivity using fMRI, with an experiment demonstrating that network mapping can reveal the areas and systems involved in economic value representation more accurately than univariate correlation analysis alone, and a similar experiment involving passive olfactory stimulation, and ii) development of algorithms for automated “brain state” recognition, in an experiment wherein a robotic arm driven by a machine-vision system performed specific operations determined through actions imagined by a participant undergoing fMRI.
He has sat on the science committee of the “Start-up coma research centre (CRC)” project funded by Regione Lombardia, within which he co-ordinated a specific work-package on the study of resting-state functional connectivity in patients with disorders of consciousness, particularly driving the development of dedicated data preprocessing pipelines for motion artefact reduction in resting-state fMRI. He also worked on quantitative in-vivo and ex-vivo imaging projects, co-authored a review on clinical applications of functional connectivity mapping and another on brain connectivity and autism. Throughout the first period of this appointment, before his relocation to the Brighton & Sussex Medical School and change to external collaborator, he co-ordinated a partially-independent research group with dedicated resources and facilities, including a cognitive neuroscientist, an engineer and a biologist
Full-time research fellow, grade 7
Throughout this appointment in the team of Prof. Hugo D. Critchley, he conducted and supported research projects on multimodal functional neuroimaging, combining fMRI, ERPs, functional near-infrared spectroscopy (fNIRS) and autonomic response monitoring. Main research topics included: i) reviewing literature on the implementation of physiological monitoring devices during MRI, particularly associated signal processing and safety aspects, ii) conducting a study on emotional modulation of visual cortex responses using fNIRS, iii) investigating the physiological contamination of the fNIRS signal by systemic factors and the correspondence between the fNIRS and fMRI signals, with development of custom hardware and mechanics, iv) studying the effect of cardiac cycle and baroreceptor activity on nociceptive stimulation ERPs, v) brain imaging in high-functioning autism, particularly brain connectivity and sexual dimorphism in Aspergers’ syndrome, language and spatial processing and large-scale grey matter texture, vi) associative memory and implicit learning as a means to assess consciousness, vii) decision-making processes under time pressure, viii) somato-sensory perception and affective touch, ix) gustatory synaesthesia and association between joint hypermobility, anxiety and brain structure. In the latter part of this appointment, he commenced his doctoral research which was later completed at the Fondazione Istituto Neurologico Carlo Besta (see above).
Feb 2005 - Jan 2008 – Freelance consultant
Feb 2008 - Jan 2009 – Collaborator, full-time research fellow
He conducted and participated in research projects on multimodal MRI, bio-signal and image analysis with the Scientific Directorate and Neuroradiology Units. For the initial part of the appointment, topics mainly included: i) developing a tailored imaging protocol to study substantia nigra degeneration in Parkinson’s disease, ii) reviewing literature on normal brain ageing investigation using MR 1H-spectroscopy, diffusion and functional MRI and on Alzheimer’s disease, iii) diffusion-tensor imaging (DTI), increasing sensitivity to degenerative change through adoption of higher-rank tensors capable of capturing complex fibre crossing architecture (generalized DTI) and through removal of the assumption of Gaussian diffusion using bi-exponential and diffusional kurtosis models, iv) use of artificial neural networks to rapidly generate diffusional parametric maps, v) reviewing literature on the physical foundations of diffusion imaging and related mathematical methods, vi) conducting a normative study of quantitative brain 1H-spectroscopy. For the second part, topics mainly included: i) fMRI on plasticity in malformations of cortical development, ii) ERP study of dorsal and ventral visual pathway engagement in reading, iii) spatial correspondence between fMRI activations and cortical current density maps derived from low-resolution ERPs, iv) combined fMRI/ERP studies on musical syntax processing and sensory consonance, v) fMRI studies on memory lateralization, resting-state functional connectivity and myoclonus epilepsy.
Consulting activities in the field information technology and microelectronics with provision of design support, co-ordination and management of research and development projects in the areas of biomedical telemetry (Microsystems Srl), systems electronics (MS Lab Srl), machine vision (NeuriCam SpA).
In the context of these activities he worked alongside senior engineers and acquired project management skills as well as proficiency in the Assembly, C and C++ languages under Linux and UNIX, in analog and digital electronic design using the EdWin, OrCAD and TannerTools packages, and in testing and characterization of electronic device prototypes.
In particular he was involved in the development of the hardware and low-level software of an early smart camera based on the ADSP-2181 processor and a CMOS sensor, supported a variety of FPGA designs and design revisions for a CMOS image sensor design (0.35 um), co-ordinated the development of a biomedical telemetry device base on GPRS (WAD/2) and worked on a high-speed optical cheque reading and sorting system. In the same context he developed a hexapod robot (GOLEM-1) based on a dedicated tree computing architecture hinged around the TOTEM neuro-chip.
Company co-founder and board member.