Travel grant reports 2015
Daniel Margulies, Max Planck Institute for Human Cognitive and Brain Sciences, Germany
Purpose: Scientific meeting, Brainhack Americas in Queretaro, Mexico
Brainhack provides an opportunity to collaborate on open science projects across institutional boundaries. Attendees come from a range of disciplines and skill levels, and have the opportunity to share knowledge and interests over a three-day period. This most recent event, Brainhack AMX, was held simultaneously at sites throughout the Americas, of which I participated in the flagship event in Queretaro, Mexico.
During the Brainhack, I developed the Python package 'SurfDist', which enables the calculation of geodesic distance on the cortical surface, while excluding topological confounds, such as the medial wall. It is also possible to use FreeSurfer labels as region-of-interest inputs for the distance calculation. The package is openly available at www.github.com/margulies/surfdist, and has recently been submitted to the post-conference proceedings special issue described below.
I will be further helping to enable knowledge transfer from Brainhack AMX by co-editing post-conference proceedings, which will be published through the open-access journal GigaScience. The aim of this novel publication format is to provide a means for attendees at Brainhack, a conference structured around participant-generated content and collaborations, to report on their achievements through a peer-reviewed publication.
Maria Angeles Correas Marin, University Complutense of Madrid, Spain
Purpose: Research visit to the SaptioTemporal Brain Imaging Laboratory at the San Diego State University (SDSU)
The main aim of the visit in SDSU was to expand my knowledge about the integration of neuroimaging techniques in order to study the change in brain dynamics. The host laboratory applies cuttingedge neuroimaging methods to study the effect of alcohol consumption during the performance of different cognitive paradigms. As this approach is closely aligned with my dissertation topic, it has been a very useful opportunity to study the brain activity of different populations of alcohol consumers during different cognitive paradigms. During the research visit I have had the opportunity to share daily meetings with Dra. Ksenija Marinkovic and also I have attended to weekly roundtables and meetings organized by the Department in a rich multidisciplinary environment that includes MRI engineers, computational neurologist, psychiatrist and neurologist. In conclusion, during these months I have improve my capabilities of studying the brain networks through different neuroimaging approaches such as Magnetoencephalography (MEG), Electroencephalography (EEG) and Magnetic Resonance Imaging (MRI). I have focus in the combination of dynamics and structural brain information, with the aim of characterizing the underlying changes in the brain functional organization, induced by compulsive alcohol consumption during adolescence. I would like to express my sincere gratitude to the INCF, as well as to the SpatioTemporal Brain Imaging Laboratory at the SDSU for making possible and for supporting my visit research.
Jules Lallouette, Inria, the French National Institute for computer science and applied mathematics, Lyon-France
Purpose: Research visit to Department of Electronics and Communication Engineering and BioMediTech Tampere University of Technology, Tampere, Finland. September 30th - October 4th 2015.
Thanks to the INCF, I was able to visit Jari Hyttinen's group in Tampere University of Technology. We worked on a joint paper on modeling mixed neuron-astrocyte networks. The model considered in this paper is based on both the INEX neuronal network model developed by Kerstin Lenk and on the simplified UAR model of astrocyte calcium activity that I developed during my Phd. Thanks to the work of Eero Räisänen, the model is working and already produced most of the results needed in the paper, we thus aim at publicating in the near future. I had the opportunity to get familiar with the calcium signaling modeling that is done in Jari Hyttinen's group, including modeling of intracellular calcium activity in astrocytes. Since I am also interested in calcium activity in astrocyte processes, we might further collaborate on this topic. In addition, we discussed the type of experimental setups that they had access to in order to acquire data on mixed neuron-astrocyte networks. We then visited the laboratory of biologist collaborators in which Micro Electrode Array experiments with mixed neuron-astrocyte cultures are being conducted. We discussed about implementation of large scale astrocyte networks and I gave a technical presentation of the inner workings of the current version of our simulator. Finally, I gave a presentation of my Phd work on astrocyte network modeling to the team, discussing mostly the influence of network topology on calcium dynamics in astrocyte networks.
Fikret Emre Kapucu, Tampere University of Technology, BioMediTech, Tampere, Finland
Purpose: Participating the IEEE EMBC 2015 conference and workshop "A Hands-on Approach to Neural Connectivity Inference Methods" where key speakers are Luiz A. Baccalá and Koichi Sameshima.
I planned to have discussions on the applicability of their suggested multivariate connectivity analysis methods on the data I am experiencing and also discussions about practical matters which can come out during the data analysis. Additionally, I intended to share my own work with the community as a presenter during oral and poster sessions and search for different viewpoints on my analysis methods. I presented the accepted articles IEEE EMBC 2015 conference: "A Fast Stimulability Screening Protocol for Neuronal Cultures on Microelectrode Arrays" and "Quantifying in Vivo and in Vitro Neuronal Bursts by an Automatized Adaptive Burst Detection Algorithm and Entropy Measures". During these presentations some interest raised on my open source tools and methods as well as the data I have used to develop these algorithms. I had fruitful discussions to develop my studies further. I also attended the full day workshop "A Hands-on Approach to Neural Connectivity Inference Methods". The presenters of the session provide their open source methods. I had discussions with Prof. Baccalá and Prof. Sameshima as well as with Dr. Laura Astolfi and Lucas Massaroppe during the session. I had the opportunity to test their methods. I have obtained valuable ideas for the interpretation of my own algorithm which is under development and also practical information for the use of their suggested method with the data recently I am working.
I would like to express my gratitude to the INCF for supporting my travel.
María del Carmen Martín-Buro, Universidad Politécnica de Madrid, Spain
Purpose: Research visit to the Episodic Memory Lab at the University of Birmingham for knowledge transfer in the study of human behaviour and brain oscillations during memory processing. 17th August- 17th October, 2015.
The purpose of the visit was to have some hands-on training in brain oscillations analysis while working on the design of a memory paradigm for a number of forthcoming neuroimaging studies. During this time I have been directly collaborating with Dr. Bernhard Staresina in three principal lines: (i) the development of a novel paradigm to assess memory consolidation, (ii) the analysis of a Magnetoencephalography (MEG) dataset, and (iii) the understanding of the role of oscillations in the cognitive processing. I have also been integrated in a wide and outstanding group of PhD and Postdoc researchers in the field of human cognition and neural oscillations, led by other experts of the School of Psychology such as Dr. Maria Wimber and Dr. Simon Hanslmayr. Thanks to this close collaboration I could attend to the recent set-up of a new Eye Tracker in the School of Psychology in Birmingham and visit the MEG system at the University of Nottingham (UK). There, I had the opportunity to observe the features of its hardware and software and learn how they prepare the participants and record the signals.
Regarding the first point, I conducted a pilot study to refine the new memory task, made changes and presented results for discussion in order to find the behavioral effect for the next neuroimaging studies. I have gained better understanding of event-related designs both in MEG and functional Magnetic Resonance Imaging (fMRI); this is technically important given the different nature and dynamics between the magnetic and BOLD activity. Moreover, I have used in-house MATLAB scripts with the software Psychtoolbox, a useful and widespread toolbox to present stimuli and record participant’s responses accurately.
In second place, the stepwise analysis of the MEG dataset comprised: preprocessing, event-related fields (ERFs) and time-frequency analysis (TFRs). I have gained experience in visual artefact detection and rejection by means of independent component analysis. During the next weeks we will extended the analysis to the ERFs and TFRs. My new understanding in these analyses allowed me to perform them on some of our MEG and iEEG datasets in the Center of Biomedical Technology in Spain.
In third place, I attended several meetings, journal clubs and seminars inside the School of Psychology and I had the opportunity to discuss and share ideas with PhD students and Postdoc researchers. This stimulating environment allowed me to gain deeper understanding about brain oscillations and cognition. My discussions with my supervisor and the rest of the Memory and Learning Group have really guided me into the best direction of task designing and data analysis. The new framework, the hypothesis-driven approach and technical skills I acquired have set the foundations of my PhD and I believe this visit means the inception of a strong collaboration in the future.
I want to express my gratitude to INCF for the travel grant and for the University of Birmingham for supporting my visit.
Shumail Mohy ud Din, National University of Science & Technology, Pakistan
Purpose: Visit and Summer Internship at German Neuroinformatics Node (INCF Germany Node), led by PD Dr. Thomas Wachtler
The purpose of visit was to spend summers productively with scientists at German Neuroinformatics Node(G-Node) and have some hands-on training in Neuroinformatics while working on on-going projects actively alongside scientists at G-Node. G-Node had planned my internship in 2 phases. In First phase, I worked over improving the G-Node Conference App for android. It had serious issues and used to take long setup time which I re-programmed and brought the setup time to less than few seconds. The App was then used for INCF Congress 2015, that took place in Australia in August and Bernstein Conference 2015 as well and will be used for conferences to happen in future as well.
In 2nd phase, I implemented a module in IRIS (https://github.com/wachtlerlab/iris) to calibrate camera for transformation of RGB to LMS. In neuroscience, at many times, research requires carefully calibrated images that accurately represent the light that reaches the camera sensor. For example, to address questions about the early visual system, images should not deviate systematically from the patterns of light incident onto, and encoded by, the retina. But due to limitations like - photography with different cameras, different lenses or focal lengths, or using lossy algorithms for compression, such research involving natural images thus relies on well-calibrated images and obtaining well-calibrated images require well calibrated camera. In my time there, I implemented 3 new tools called LED-Pseudo-Monochromator, LED-Photo-Spectrum and LED-PWM-Thresholder in IRIS which help in process to calibrating camera.
Besides this, the arranged training sessions and seminars covered different areas including, but not limited to, Efficient coding with GIT, The life and death of program functions, Introduction to Multi-threading and parallel programming, Quantification of Spatial resolution(acoustically guided predators), Efficient coding of a dynamic trajectory by grid cells in the entorhinal cortex, High Resolution 3D tracking for quantitative analysis of exploratory behavior, Unsupervised neural spike identification for large-scale, high-density micro-electrode arrays,
Thank you German Neuroinformatics Node and INCF for providing this opportunity and facilitating with INCF Travel Grant, which made this possible.
Natalia Díaz Rodríguez, Åbo Akademi University, Finland
Purpose: Research collaboration with Prof. Lise Getoor at the LINQS (Statistical Learning and Reasoning) Group at the Computer Science Department of University of California, Santa Cruz for extension of PhD topic on human behavior modelling and recognition for Ambient Assisted Living to include statistical learning with Probabilistic Soft Logic framework.
Thanks to INFC, I was able to visit one of the most cross-disciplinary universities, located in the beautiful coastline of north California. The group has developed a versatile and interpretable first order logic tool for modelling problems from image segmentation, stance-detection, sentiment analysis, drug target prediction, latent social groups and trust, ontology alignment, recommender systems, knowledge graph extraction, document classification, spammer detection, etc. The tool is Probabilistic Soft Logic (PSL), an open source probabilistic programming language that supports declarative features, collective reasoning and lifted models (which capture common, repeated patterns of dependencies, and serve to define semantics using fewer parameters, better learning and generalization, and faster –convex optimization- inference when exploiting commonalities). I applied PSL’s data-driven approach to learn relational dependencies, cycles and weight importance in uncertain data, by adapting the Cornell Activity Dataset within the problem of human activity recognition considering fuzzy and probabilistic uncertainty.
As a side task, together with Sabina Tomkins, we are organizing Data Science Social Series for the fall 2015 at CS Dept of UC Santa Cruz, e.g. a data science and digital humanities mixer for Humanities and Science. Among others, we are arranging a panel on recent advances on computational creativity and algorithmic evaluation of art with Deep Learning next 19th Oct 2015.
Thanks to being visiting research scholar at UCSC, I could attend the Summer Institute of Cognitive Neuroscience in Santa Barbara for two weeks. I learnt about neuroscience as a science that, together with engineering, psychology, medicine and physics, adds a crucial intertwining element when it comes to modelling the brain, and serves as fascinating inspiration to model artificial intelligence from natural intelligence.
I gave a guest talk about my PhD research at Stanford university, hosted by Prof. Scott Delp’s group with Jen Hicks and Joy P. Ku, and initiated collaboration among them and Philips Research Eindhoven. Also delivered the talk (snippet on YouTube) at General Assembly San Francisco within Product Management Course (led by Thabet Alfishawi, Product Manager at Youtube) where I met the most interrupting, engaged and motivated students (about 5 linkedin/twitter invitations within the next 24h).
I got a grant from NSF to attend the NSF Data Science workshop with cross-disciplinary scientists where we discussed main problems, different and similar methods across fields, and we are working together on a paper on a jargon-free data-science mission.
I attended the Data Science Summit+ Dato Conference in San Francisco in July’15, where top academics and industry speakers talked about latest trends on machine and deep learning. We initiated some collaboration with other researchers, such as Amanda Casari, Tsvi Achler. Also made collaborator connections at the Data Science Handbook fireside chat event, e.g., Luis M. Sanchez (CEO at Ttwick and computational creativity scientist for algorithmic music) which helped us with the panel organization.
I attended DevelopHer Hackathon at LinkedIn headquarters in Sunnyvale and developed a site for confidential 1-1 mentorship: Tiramisu: Bringing you up through bytesize coaching (5’ pitch on YouTube). Also visited Baidu offices and learnt more about fascinating current advances on Deep Speech.
Arvind Kumar, KTH Royal Institute of Technology, Sweden
Purpose: To participate in a workshop organised by Gaute Einevoll on the topic of 'LFP modelling and analysis' in Oslo, and to visit the labs of Dr. Torkel Hafting-Fyhn and Dr. Marianne Fyhn at the Oslo university.
The discussions at the workshop precipitated in defining a collaborative project between Gaute Einevoll’s group and mine. We will investigate the contribution of the local spiking activity in the striatum to the local-field-potential (LFP) measured in the striatum. I will provide spiking activity dynamics from large- scale models of the striatum1 and we will use the LFPy tool2 developed at the Einevoll lab to generate the LFP signals. Prof. Hellgren-Kotaleski from KTH, will provide the detailed neuron model of the medium spiny cell that is needed to simulate LFP using the LFPy tool. In addition, in discussion with Torkel and Marianne we have identified at least two domains of common interest. The first idea relates to the detailed examination of the grid cell data to extract the trial-by-trial variability information. The second idea relates to testing of one our model of oscillations in which oscilla- tory activity arises due to the imbalance of excitation and inhibition3 and not so much due to the increase in synaptic weights. This was the first interaction with the Hafting-Fyhn group and we have continued our discussions after my return. Soon I plan to host them at KTH, Stockholm. We are positive that given many common interests we will have a collaboration in the near future. So all in all this was a very successful visit. And I am very thankful to the INCF for providing the travel funding that made this visit possible.
Ilaria Colombi, Italian Institute of Technology (IIT), Italy
Purpose: Visit and collaboration with the laboratory of Theoretical Neurobiology & Neuroengineering lead by Prof. Michele Giugliano at the University of Antwerp, Belgium
Thanks to INCF I had the possibility to visit and start an active collaboration with the laboratory of Theoretical Neurobiology & Neuroengineering lead by Prof. Michele Giugliano at the University of Antwerp (Belgium) for three weeks (i.e. 3/07/2015 and 24/07/2015). The aims of my visit were to learn new methodologies to analyze and characterize the features of neuronal network dynamics, and to apply those analysis tools to electrical signals from acute rat brain cortical slices coupled to MEAs upon pharmacological and ionic manipulation. We recorded two type of brain slices: the hippocampus-entorhinal cortex (EC) and the neocortical slices with perforated and 3D Arrays of substrate-integrated MicroElectrodes (MEAs). Thank to my previous knowledge on the signals exhibited by the cortical cultures, I could adapt my MATLAB scripts for the analysis of the newly acquired data. During my staying in Antwerp, I had the possibility to characterize the signal of acute rat brain slices both at the local field potential (1-300 Hz) and multiunit activity level (300-3000 Hz), in control condition and during pharmacological and ionic manipulation. Thanks to INCF I had my first experience abroad, where I could interact with an enthusiastic team made of top-quality scientists from many diverse disciplines. Moreover, during my stay, I had the possibility to do a presentation on my master thesis results and their feedback will be very useful to re-examine my data under a different perspective. The new technical skills that I have acquired will be fundamental for my PhD training, that I will start in November. In particular the novel analytical approach developed here will, in turn, help us answering key questions on network pathophysiology (e.g. epilepsy and stroke), for the development of innovative electro-chemical therapies. I believe that this training not only could empower my career development, but will also provide me with a novel framework to unleash my full potentials.
Azeez Ayodeji Adebimp, Université de Picardie Jules Verne, Amiens, France
Purpose: Research visit to University of Groningen, The Netherlands, 27th April to 13th May, 2015 for knowledge transfer on NIRS and multichannel EEG data analysis
Host: Prof. Maurits Natasha, UMCG, Gronigen
The purpose of the visit is on the exchange of knowledge on NIRS and multichannel EEG data analysis particularly focusing on brain connectivity in neonates. It is expected to give opportunity to learn more about neuroscience, experimental settings, programming and computational modeling skills.
Activities: I was situated at the Neuroimaging centre at the university medical centre under the supervision of Dr Branislava Ćurčić-Blake, that taught me the dynamic casual modelling, a form of effective connectivity. I analysed some of their experimental data and presented results for discussion. I visited the neonatology department to learn measurement techniques and analysis methods of neonate data. The MATLAB dependent software built by the department was demonstrated to me as I am also developing specific software for neonate data analysis. I met several professors and technicians working on research that relate to my thesis for discussion and seek for their opinion and advice. I attended several classes, journal clubs and public lectures inside the University Medical Centre, as well as journal classes. I had discussion with some of the PhD students and share ideas with them. Also, I showed some of their master students the basic processing of EEG data.
Accomplishments: I was able to understand the dynamic causal modelling. I applied the method on EEG data to measure the influence of some features in neonate brain development. I presented the results to the centre on the last day before leaving and we agreed to improve the results and publish it in a journal. My discussions with some of the professors have really help me to the best direction of data analysis. This visit has created a kind of collaboration for the future. I am very grateful to INCF for the travel grant and will surely include INCF in the part of accomplishment in the journal article to be published.
Jose Angel Pineda-Pardo, Universidad Politécnica de Madrid, Spain
Purpose: Research visit to the Center for Neuroimaging and the Indiana Alzheimer Disease Center at Indiana University for knowledge transfer in the study of human brain connectivity, 1-30 April 2015.
The main aim to the visit in the IU was to collaborate with PhD Joaquín Goñi in the development of strategies of analysis in human brain structural and functional networks. During this time we have collaborated in three principal lines, 1. the development of a novel pipeline of functional and structural connectivity analysis using functional MRI (fMRI) and diffusion weighted imaging (DWI), 2. the study of functional networks in a longitudinal cohort of healthy young subjects that underwent cognitive training, 3. the study of functional networks in a local cohort of mild cognitive impairment (MCI) subjects. We also had several meetings with other experts in Bloomington Indiana University and in Eli Lilly’s company, which might probably represent the seed for future collaborations.
Concerning the first point, the pipeline of functional connectivity analysis using fMRI is now fully operative, and has been applied to the cognitive training dataset. Using a multiplex network partitioning approach we generated some preliminary results, which will be extended in the next weeks. The local cohort of MCI is now being processed, and we expect to generate very interesting answers in the possible links between brain functional connectivity, and other described biomarkers such as amyloid plaques accumulation, cognitive decline, risk genetics groups or brain atrophy.
Acknowledgment: I would like to express my gratitude to the INCF, as well as to the Center for Neuroimaging and the Indiana Alzheimer Disease Center at Indiana University for supporting my visit.
Kerstin Lenk, Tampere University of Technology, BioMediTech, Finland
Purpose: Research visit to laboratory of Theoretical Neurobiology and Neuroengineering at University of Antwerp for knowledge transfer regarding computational models and multi-electrode array experiments, 12-15 April 2015
This research visit to the laboratory of Theoretical Neurobiology and Neuroengineering at University of Antwerp, led by Prof. Michele Giugliano, produced the following results:
1) Knowledge transfer: Our PhD student Inkeri Vornanen and I gave each a talk about our current projects in Computational Neuroscience. This led to a very fruitful discussion about dynamics and topology in neuronal networks. Michele Giugliano and his lab members presented ongoing multi-electrode array (MEA) experiments with acute rat brain cortical slices and how stimulations using optogenetics are conducted. Furthermore, we discussed other MEA types which are or will be applied in one of the labs. We exchanged insights about several simulation and analysis tools.
2) Research visit of Inkeri Vornanen: We defined the details of a 3-month research stay of PhD student Inkeri Vornanen in Michele Giugliano’s lab. She is currently supervised by Prof. Jari Hyttinen and me, both from Tampere University of Technology, Finland. The research topic of Inkeri’s stay will be stimulation of neuronal networks on MEAs, where she will get a detailed introduction during her research stay, and the simulation of those networks. We aim for a high-impact journal publication.
3) Grant applications: We discussed several national and international grant applications which we will start to formulate soon.
Acknowledgment: I would like to express my gratitude to the INCF for supporting my visit.
Paulo de Castro Aguiar, INEB - Instituto de Engenharia Biomédica, Portugal
Purpose: Research visit to TNB - Theoretical Neurobiology & Neuroengineering Lab, U. Antwerp, for knowledge transfer regarding multi-electrode arrays (MEAs), 8th - 21st of February, 2015
This research visit to the Theoretical Neurobiology & Neuroengineering Lab, U. Antwerp, led by Prof. Michele Giugliano, allowed the following accomplishments:
1) Knowledge transfer. During the research visit it was possible to be actively engaged with ongoing experiments using in vitro multi-electrode arrays. This allowed a better understanding of methodologies and techniques associated with this technology, in experiments involving either cell cultures or acute brain slices. Particularly, in terms of data analysis, it was possible to appreciate a different paradigm of data organization/analysis: MEAs produce vast amounts of data, and the strategies used in Michele Giugliano’s lab to address this situation are insightful and very useful.
2) Fostering collaboration. This visit allowed the identification of areas of common interest and complementary expertise between PAguiar and MGiugliano. Work started to be developed in two areas: connectivity inference in the context of in vitro MEAs data, and simulations of network dynamics associated with experiments combining MEAs and optogenetics. Key open questions were identified and some strategies to address them were discussed. This work continues to be developed after the visit.
3) Research grant submission preparation. Taking advantage of common research interests, this visit was also used to start defining a research project proposal which will be submitted in a near H2020 call.