Department of Biomedical Engineering and Computational Science

BECS Coffee Seminar

The BECS Coffee Seminar showcases contemporary research at BECS to catalyze collaboration among the researchers. It also serves as a venue for practicing, e.g., master's thesis and conference presentations and as a forum for talks by visiting researchers.

The seminar is organized every Wednesday (except when there's overlapping activity) at 2:15 PM (coffee and biscuits served at 2:00 PM) in the BECS Coffee Room (F336). The talks usually last about 15 minutes.

The Coffee Seminar is organized by Dmitry Smirnov, Pietro della Briotta Parolo and Julio Hernandez Pavon. All researchers with potential talks are encouraged to contact them.

Seminar schedule, 2014

Date: Wednesday, December 10, 2014
Speaker: Mikael Eriksson (Philips Medical Systems MR, Finland)
Topic: Comparison of five methods for deformable, multi-modal image registration in prostate and pelvic area

Abstract: Radiation therapy (RT) planning in prostate cancer typically uses both CT images and T2-weighted MR images, for dose calculations and treatment volume delineation, respectively. By using MRI to generate synthetic attenuation maps based on a T1-weighted, double-echo, gradient echo sequence with Dixon reconstruction (mDixon), CT imaging can be discarded from the planning process, and dose delivery uncertainties due to e.g. patient positioning and image registration errors decrease. Multi-modal image registration may still be needed to ensure that the mDixon and T2-weighted images are aligned. The aim of this thesis was to compare five deformable methods for this registration problem; the Demons algorithm, B-spline registration, a fast elastic image registration (FEIR) method, a locally affine hierarchical and a locally rigid hierarchical method. The methods were evaluated qualitatively (visual assessment) and quantitatively for 19 cases, using anatomical landmark accuracy, normalised cross-correlation of gradient images and the Dice similarity coefficient (DSC) of the bladder, prostate and seminal vesicles as evaluation metrics.

The results showed that in cases following the normal RT clinical work-flow, image alignment was good even without registration, but was improved with registration. When intentional patient movement impaired the initial alignment, the benefit of image registration increased. FEIR was the best method in terms of visual assessment, quantitative evaluation performance and computational speed, and seems well-suited for use in the MR-only radiotherapy work-flow for prostate cancer.

Keywords: Deformable image registration, MR-only radiation therapy, prostate cancer

Date: Wednesday, October 29, 2014
Speaker: Andrea Lancichinetti (Umeå University, Sweden)
Topic: Richer models for complex networks

Abstract: Random walks on networks is the standard tool for modelling spreading processes in social and biological systems. However, this conventional first-order Markov approach ignores a potentially important feature of the dynamics: where flow moves to may depend on where it comes from. Here we analyse multi-step pathways from different systems and show that ignoring the effects of second-order Markov dynamics has important consequences for community detection and ranking, but only marginal consequences for disease spreading through air travel. When people travelling by plane can transmit infections to random other people in cities, the memory effects from their travel patterns are lost. Therefore, accurately modelling air travel patterns has a negligible effect on disease spread. However, when focusing on the travel patterns themselves, or on systems with limited mixing, we observed that random flow on networks understates the effect of communities and exaggerates the effect of highly connected nodes. For example, capturing dynamics with a second-order Markov model allows us to differentiate airport hubs from popular destinations and reveal actual travel patterns in air traffic, and to uncover multidisciplinary journals and ranking that favour specialized journals in scientific communication. These findings were achieved only by using more available data and making no additional assumptions, and therefore suggest that accounting for higher-order memory in network flows can help us better understand how real systems are organized and function.

Andrea Lancichinetti is a postdoc researcher at Icelab, Umeå University. His research is about complex networks with a special focus on clustering algorithms and models of the structure and network flow.

Date: Wednesday, May 28, 2014
Speaker: Prof. Petter Holme (Umeå University, Sweden, and Sungkyunkwan University, South Korea)
Topic: The Social, Economic and Sexual Networks of Prostitution

Abstract: Prostitution is, on one hand, a unique socioeconomic phenomenon, on the other hand, it can tell us much about society as a whole. Selling sex has rarely been accepted socially; it has often been prohibited and suppressed; still it exists in almost all the World’s countries, and has done so for thousands of years [1]. Furthermore, prostitution is female dominated, labor intensive, requires little education, and still it is well paid. This contradicts mainstream theories of labor economics [2]. Moreover, sex-buyers are almost exclusively male, but apart from that, they have few traits in common. They are very uniformly distributed among different age classes, professions and ethnicities [3]. These examples illustrate the special nature of prostitution as a socioeconomic system. One may wonder if prostitution can give us any information at all about society as a whole? Is it an isolated system with a unique set of rules and mechanisms? Yet more peculiarities are that the social norms around prostitution vary much between different societies. The same is true for the legal statuses that range from legal and regulated to laws against both buying and selling sex. Economists are also interested in the unusual age–wage curve of sex sellers (leveling around 30–40 years to decrease for yet older sellers) [4].

The fact that prostitution has to circumvent the mainstream channels of advertisement creates a more laterally organized economy. Not unlike the new network economy driven by social media. We argue that, even though most research questions about prostitution are rather specific, it can also teach us something more general about network-economic systems emerging today. To reach these conclusions, we review prostitution research with a focus on Internet mediated forms. In particular, we will study many aspects of a dataset derived from an online community serving sex-buyers needs [5,6]. At this community, the users log and review their sex purchases, giving the date and time of the encounter as well as a rating of the experience. We use this data to study online-offline feedback effects over a sex-sellers time of presence in the data as well as spatial factors. We briefly mention (with the above-mentioned dataset as example), the use of this type of dataset for the study of infectious disease spreading [7].

[1] Davis, K. 1937. The Sociology of Prostitution. Am. Soc. Rev. 2, 744–755.
[2] Edlund, L. and Korn, E. 2004. A Theory of Prostitution. J. Pol. Econ. 110, 181–214.
[3] Pitts, M. K., Smith, A. M. A., Grierson, J., O’Brien, M. and Misson, S. 2004. Who pays for sex and why? An analysis of social and motivational factors associated with male clients of sex workers.’ Archives of Sexual Behavior 33, 353–358.
[4] Edlund, L., Engelberg, J. and Parsons, C. A. 2009. The wages of sin. Columbia University Economics Discussion Paper, 0809–16.
[5] Rocha, L. E. C., Liljeros, F. and Holme, P. 2010. Information dynamics shape the sexual networks of internet-mediated prostitution. Proc. Natl. Acad. Sci. USA 107, 5706–5711.
[6] Holme, P. 2012. Social, sexual and economic networks of prostitution. Leonardo 45, 80-81.
[7] Rocha, L. E. C., Liljeros, F. and Holme, P. 2011. Simulated epidemics in an empirical spatiotemporal network of 50,185 sexual contacts. PLoS Comp. Biol. 7, e1001109.

Date: Wednesday, March 26, 2014
Speaker: Prof. Michael D. Noseworthy, Ph.D., Co-Director, McMaster School of Biomedical Engineering, Department of Electrical and Computer Engineering, McMaster University.
Topic: Methodological Considerations for Brain Magnetic Resonance Spectroscopy of the Brain

Abstract: Magnetic resonance imaging (MRI) has it's roots in nuclear magnetic resonance (NMR). With human MRI systems NMR spectroscopy is simply referred to as magnetic resonance spectroscopy (MRS). Routine brain MRS involves assessment of mobile proton containing metabolites. The technique is plagued by numerous overlapping resonances in the spectral frequency domain. Yet there is still valuable information to be gained. The MRS method however is more versatile than simply assessment of protons. Nuclei possessing the quantum mechanical property of spin can in theory be probed. In this presentation I will describe brain proton MRS and some of the recent advances we have made. In addition I will describe methodological considerations around application of other physiologically important nuclei, namely 13C, 23Na and 31P, in the understanding of human brain health and disease.

Short Bio: Dr. Michael D. Noseworthy, PhD, PEng. Received a M.Sc. from the University of Guelph for work in the evaluation of anaesthetic hepatotoxicity using MRI and in vivo 31P-NMR. He then obtained a PhD from University of Guelph (1997) specializing in applications of MRI/MRS and electron paramagnetic resonance (EPR) methods to assess free radical induced brain damage. From 1997-1999 Dr. Noseworthy was a postdoctoral fellow in Imaging Physics, Sunnybrook Health Sciences Centre working on the evaluation of cancer microvasculature through development of correlative MRI and energy dispersive X-ray microanalysis (EDXS). he had his first staff position as a MRI physicist at The Hospital for Sick Children, Toronto, and Assistant Professor in Medical Biophysics and Medical Imaging, University of Toronto, starting in 2000. He was recruited by Dr. John Bienenstock to St. Joseph’s Healthcare and Brain-Body Institute, McMaster University in August 2003. Following 3 years as an Assistant Professor in Radiology and Medical Physics at McMaster University, Dr. Noseworthy was appointed to Electrical & Computer Engineering at McMaster University, where he currently resides as an associate professor. Also at McMaster University he is the Co-Director of the School of Biomedical Engineering and is the Scientific Director of the Imaging Research Centre, at St. Joseph’s Healthcare, Hamilton. Dr. Noseworthy is a professional engineer (biomedical) and member of the Professional Engineers of Ontario (PEO) and the Institute of Electrical and Electronics Engineers (IEEE). He has published over 250 peer reviewed papers and conference proceedings, given almost 100 invited lectures world wide, and trained 39 trainees in his lab. He has numerous awards including a recent McMaster President’s Award for Excellence in Graduate Teaching. His research interest is the assessment of tissue microstructure and metabolism using magnetic resonance imaging (MRI) and multinuclear in vivo nuclear magnetic resonance (NMR) spectroscopy methods. In his spare time Dr. Noseworthy is an avid ice hockey player and coach.

Seminar schedule, fall 2013

Date: Wednesday, December 18, 2013
Speaker: Dr. Timo Nyberg (BIT Research Centre)
Topic: Intelligent traffic systems

Abstract: Intelligent traffic systems are revolutionizing the transport in the modern society. Transportation means will be based more on services instead of owning e.g. your own cars. The enablers for this transition are the mobile internet, smart phones and big data based service offering. One interesting area for research is the optimization of predictive dynamic routing where several uncertainties like weather and unexpected accidents are inherent.

Date: Wednesday, November 27, 2013
Speaker: Prof. Marko Turpeinen (Director of Helsinki EIT ICT Labs)
Topic: EIT ICT Labs - Bringing Innovations to Life

Abstract: EIT ICT Labs is one of the first Knowledge and Innovation Communities (KIC) set up by the European Institute of Innovation and Technology (EIT), as an initiative of the European Union. EIT ICT Labs' mission is to drive European leadership in ICT innovation for economic growth and quality of life. Since 2010, EIT ICT Labs has consistently brought together researchers, academics and business people. By linking education, research and business, EIT ICT Labs empowers ICT top talents for the future and brings ICT innovations to life. EIT ICT Labs' partners represent global companies, leading research centres, and top ranked universities in the field of ICT. EIT ICT Labs consists of six Co-location Centres ­ Berlin, Eindhoven, Helsinki, Paris, Stockholm and Trento ­ and will turn these already excellent regional centres into word-class innovation hotspots.

Find out more about EIT ICT Labs at

Date: Tuesday, November 19, 2013
Speaker: Dr. Vadim V. Nikulin (Neurophysics Group, Department of Neurology, Charité - Universitätsmedizin Berlin, Germany)
Topic: Advances in Human EEG: recording methods, analytic techniques, and oscillatory phenomena

Abstract: In the first part of the talk I will present a new type of EEG electrodes. Current mainstream EEG electrode setups permit efficient recordings, but are often bulky and uncomfortable for the subjects. Recently we introduced a novel type of EEG electrode, which is designed for an optimal wearing comfort. This novel electrode is not felt by the subject and is close to invisible to an external observer. In the second part of the talk I will present a novel method for the reliable and fast extraction of neuronal oscillations from multi-channel EEG/MEG/LFP recordings. The method is based on a linear decomposition of recordings: it maximizes the signal power at a peak frequency while simultaneously minimizing it at the neighboring, surrounding frequency bins. Such procedure leads to the optimization of signal-to-noise ratio and allows extraction of components with a characteristic "peaky" spectral profile, which is typical for oscillatory processes. We refer to this method as spatio-spectral decomposition (SSD). In the third part of my talk I will present new oscillatory phenomenon in human EEG: monochromatic ultra-slow oscillations (MUSO). The striking signature of these oscillations is their almost monochromatic spectral profile restricted to the frequency range 0.08–0.15 Hz. We also observed a significant coherence between MUSO and Oxy-Hb (NIRS)/arterial blood pressure. We hypothesize that MUSOs can occur due to the modulation of the blood-brain barrier DC potential by mechanical fluctuations related to hemodynamics or due to the skin vasomotion. The discovery of such possibly extraneuronally produced oscillations opens an avenue for studying hemodynamic responses with EEG technology and in addition poses a number of questions concerning the interpretation of previously recorded low-frequency neuronal oscillations in humans.

Date: Monday, November 18, 2013
Speaker: Prof. Gabriel Curio (Charité - Universitätsmedizin Berlin, Germany)
Topic: TBA


Date: November 6, 2013
Speaker: Dr. Zhipei Sun
Title: Modulating light with graphene

Abstract: The richness of optical and electronic properties of graphene attracts enormous interest. For example: graphene has high mobility and optical transparency, in addition to flexibility, robustness and environmental stability [1]. Indeed, the rise of graphene in photonics and optoelectronics is shown by several recent results [1-4]. Using graphene, we generate optical pulse durations as short as 200 fs [5] capable of broadband tunability [6,7] from visible to infrared [8-10] due to the linear dispersion of the Dirac electrons in graphene [1-2]. Light-graphene interaction also can be tailored by using micro & nano-structures [11-12].

[1] F. Bonaccorso, Z. Sun Nat. Photonics 4, 611 (2010).
[2] A. Martinez, and Z. Sun, Nat. Photonics 7, 842 (2013).
[3] T. Hasan, Z. Sun Adv. Mater. 21, 3874 (2009).
[4] Z. Sun et. al. ACS Nano 4, 803 (2010).
[5] D. Popa, Z. Sun Appl. Phys. Lett. 97, 203106 (2010).
[6] Z. Sun, D. Popa Nano Res. 3, 653-660 (2010).
[7] D. Popa, Z. Sun Appl. Phys. Lett. 98, 073106 (2011).
[8] A. A. Lagatsky, Z. Sun Appl. Phys. Lett. 102, 013113 (2013).
[9] R. Mary Opt. Express 21, 7943-7950 (2013).
[10] Z. Sun Physica E 44, 1082 (2012).
[11] F. Bonaccorso, Mater. Today 15, 564 (2012).
[12] J. Mertens Nano Lett. DOI: 10.1021/nl4018463 (2013).

Date: September 18, 2013
Speaker: Prof. David Landau (Center for Simulational Physics, University of Georgia, Athens, USA)
Topic: A new paradigm for petascale Monte Carlo simulation

Abstract: While novel Monte Carlo algorithms are revolutionizing our ability to determine the properties of diverse models in statistical physics, the move to massively parallel computers is introducing new challenges to the user. After reviewing the powerful Wang-Landau sampling method that iteratively determines the density of states for a system by performing a random walk in energy space, we introduce a simple, generic, parallel Wang-Landau algorithm that is naturally suited to implementation on massively parallel, petaflop supercomputers. The new approach introduces a replica-exchange framework involving densities of states that are determined iteratively for overlapping sub-windows in energy space, each via traditional Wang-Landau sampling. The advantages and general applicability of the method are demonstrated using thousands of cores for several quite different systems (possessing either discrete or continuous degrees of freedom) including those with complex free energy landscapes and topological constraints.

Date: August 14, 2013
Speaker: Prof. Rafael Barrio (Universidad Nacional Autónoma de México, Mexico City, Mexico)
Topic: Coupled Dynamics in the Development of Stem Cells

Abstract: We shall revise a general way of treating complex systems that exhibit important dynamical processes in several time scales. We apply this concept of coevolution to the early development of stem cells in plants.

Seminar schedule, spring 2013

Date: July 31, 2013
Speaker: Prof. Alex Arenas (University Rovira i Virgili, Tarragona, Spain)
Topic: Dynamical processes in multiplex networks

Abstract: Modern theory of complex networks is facing new challenges that arise from the necessity of understanding properly the dynamical evolution of real systems. One of such open problems concerns the topological and dynamical characterization of systems made up by two or more interconnected networks. The standard approach in network modeling assumes that every edge (link) is of the same type and consequently considered at the same temporal and topological scale. This is clearly an abstraction of any real topological structure and represents either instantaneous or aggregated interactions over a certain time window. Therefore, to understand the intricate variability of real complex systems, where many different time scales and structural patterns coexist we need a new scenario, a new level of description. We present the time scales associated to diffusion processes that take place on multiplex networks, i.e. on a set of networks linked through interconnected layers. To this end, we propose the construction of a supra-Laplacian matrix, which consists of a dimensional lifting of the Laplacian matrix of each layer of the multiplex network. We use perturbative analysis to reveal analytically the structure of eigenvectors and eigenvalues of the complete network in terms of the spectral properties of the individual layers. The spectrum of the supra-Laplacian allows us to understand the physics of diffusion-like processes on top of multiplex networks.

Date: July 3, 2013
Speaker: Prof. Angel Sánchez Sánchez (Universidad Carlos III de Madrid, Spain)
Topic: Does choosing partners promote cooperation in networked Prisoner's Dilemmas?

Abstract: In the last few years, experiments from different groups have established beyond any reasonable doubt that networks do not promote cooperation when people play a Prisoner's Dilemma. Subsequently, newer experiments suggest that, if subjects can choose their partners, i.e., the network can be rewired, cooperation can be enhanced even to values close to 90% of the population. We have conducted experiments a few weeks ago to address the issue as to if and why exactly this happens. In this talk, I will present preliminary results arising from the analysis of our data that allow to understand the mechanisms at work in these situations.

Date: June 12, 2013
Speaker: Kalle Airo, Aalto Ventures Program
Topic: Aalto Entrepreneurial Ecosystem

Abstract: Aalto Ventures Program offers and develops entrepreneurial education at Aalto University. In AVP, you will learn by doing together with experienced professionals and people with various backgrounds and nationalities. This will teach you practical street smarts and entrepreneurial mindset for starting new ventures and making impact in real life. AVP works in close collaboration with Aalto's startup community (Aaltoes & Startup Sauna). In addition, AVP is being developed in collaboration with Stanford Technology Ventures Program.

Date: May 15, 2013
Speaker: Dr. Fanny Lachat & Dmitry Smirnov (BECS)
Topic: Joint attention, Inter-subject decoding of action/observation in the brain

Abstract: Interaction between individuals is one of the corner stones of human behavior. We convey our thoughts, ideas, intentions and emotions in various ways, e.g. gestures, facial and vocal expressions, music, actions, just to name a few. Common coding theory suggests that action and perception of action (observation) share their representations in the brain. We will briefly introduce these concepts and discuss the pseudo-hyperscanning paradigm that we plan to use in our project.

Date: May 8, 2013
Speaker: Dr. Aki Vehtari (BECS)
Topic: Bayesian modeling - Probabilistic computational science

I will briefly present some current and past projects of Bayesian modeling group. Presented applications are mostly in field of epidemiology and brain signal analysis.

Date: April 24, 2013
Speaker: Dr. Angel Sánchez
Topic: Networks do not promote cooperation among human subjects in Prisoner's Dilemmas

Abstract: A very popular explanation for the emergence of cooperation in social dilemmas is the existence of an underlying network of contacts constraining who one can interact with. The past two decades have witnessed a wealth of theoretical studies that have concluded that this "network reciprocity'' is indeed possible under a variety of circumstances. Actually, simulations indicate that heterogeneous networks should be particularly efficient in fostering cooperation in social dilemmas. These theoretical results have almost never been put through the test of experiments, and most of the few available experimental works deal with very small networks. In the last two years, we have carried out several experiments, including one involving 1229 subjects, and developed a supporting theory that allow us to rule out network reciprocity as a valid hypothesis to explain the emergence of cooperation among humans involved in a Prisoner's Dilemma. I will prove this by reporting on our experiments and, additionally, on a meta-analysis of other groups' experiments. Our results point to the fact that humans seem to disregard payoff differences with their neighbours, adopting a more reciprocal attitude (moody conditional cooperation), as the key factor. I will conclude by presenting results of a recent, lab-in-the-field experiment showing that this behavior is the same for subjects in a wide range of ages, except for children, who are significantly less cooperative, and for elderly people, who cooperate to a large extent. This indicates that networks do not promote cooperation among people of any age.

Date: April 17, 2013
Speaker: Dr. Simo Särkkä (BECS)
Topic: Multiple Target Tracking in Biometrics, Medical Imaging, and Complex Networks

Abstract: The problem of following multiple objects (airplanes, vehicles, people) using one or more sensor devices (radars, comint, video) is called multiple target tracking problem. In the presence of noise and uncertainties, the problem can be mathematically modeled as Bayesian estimation of time-varying parameters (or states) from multivariate measurement data. The general methodology for solving this kind of general time-varying inverse problems is called optimal filtering, and when the Bayesian model formulation is used, this methodology is called Bayesian filtering. In this talk, I will explain the basic principles of solving multiple target tracking problems with Bayesian filtering methods. I will also discuss applications of multiple target tracking methods to statistical estimation of large predator population in Finland, multi-dipole approach to MEG inversion, and to following of link activations in time-varying complex networks.

Date: April 10, 2013
Speaker: Dr. Ilkka Nissilä (BECS)
Topic: Technical developments in near-infrared imaging

Date: March 27, 2013
Speaker: Prof. Jari Saramäki (BECS)
Topic: Temporal networks - a gentle introduction

Seminar schedule, fall 2012

Date: December 19, 2012
Speaker: Prof. Jouko Lampinen and Anu Virtanen (BECS)
Topic: Results of the personnel survey of BECS 2012

Abstract: Prof. Jouko Lampinen will share the results of the personnel survey of BECS 2012, and Anu Virtanen will act as facilitator in analyzing the results in groups. Therefore, this is the perfect meeting to express what you like or what you don't like in BECS. Your suggestions are very important, so don't be shy, go there and express your thoughts. Your opinions are very important to improve the working atmosphere in BECS. This time the seminar will last about two hours, but we will have "Real Coffee", and some surprises :)

Date: December 18, 2012 (at 1 PM)
Speaker: Prof. Mitsugu Matsushita (Chuo University, Japan)
Topic: Statistical Features of Complex Systems toward establishing Sociological Physics

Abstract: Complex systems have recently attracted much attention, regardless of natural sciences or sociological sciences. Members constituting a complex system evolve through nonlinear interactions among each other. This means that in a complex system the multiplicative experience or, so to speak, history that any member has had produces its present characteristics. We can then anticipate the following. If attention is paid to any statistical property in any complex system, the lognormal distribution is the most natural and appropriate for the standard or "normal" statistics to look over the whole system. In fact, the lognormality emerges rather conspicuously when we examine, as familiar and typical examples of statistical aspects in complex systems, nursing-care period for the aged, populations of prefectures and municipalities and our body height and weight. Many other examples are found in nature and society. Based on these observations, we would like to discuss the possibility of sociological physics.

Date: December 12, 2012
Speaker: Prof. Morten Mørup (DTU, Denmark)
Topic: Modeling Brain Connectivity by Non-parametric Bayesian Relational Models

Abstract: Functional and diffusion magnetic resonance imaging have become key non-invasive measuring modalities in order to quantify the brains functional and structural connectivity respectively. In this it will be discussed how non-parametric Bayesian models for complex networks can be used to extract prominent patterns of brain connectivity in these functional and structural brain networks.

Date: December 5, 2012
Speaker 1/2: Airi Lampinen (Helsinki Institute for Information Technology HIIT)
Topic: Privacy as Boundary Regulation in Social Network Services

Abstract: While the primary reasons to use social network services (SNSs) are found in willingness to connect and share with others, few people wish to share everything, with everyone, all the time. This talk will shed light on how people regulate the boundaries of social interaction at different times in the networked context of various SNSs such as Facebook,, and Practices for interpersonal boundary regulation include cooperation among users and are not limited to online environments.

Date: December 5, 2012
Speaker 2/2: Suvi Silfverberg (Helsinki Institute for Information Technology HIIT)
Topic: Profile Work and Social Norms in Social Network services

Abstract: Profiles in SNSs are typically easy to acquire and quick to set up, yet maintaining one is not necessarily a simple and effortless matter for the user, especially when the publicized content is behavioral information captured by an automated sharing mechanism. With automated sharing of behavior information behaviors that used to be private are now made visible to others in real-time. This talk considers automated sharing of behavioral information in SNSs, focusing on the efforts maintaining a profile requires and the social norms that guide these efforts.

Date: December 4, 2012
Speaker: Prof. Marcello Massimini (University of Milan, Italy)
Topic: Probing consciousness with TMS/EEG

Abstract: Theoretical considerations suggest that consciousness depends on the ability of neural elements to engage in complex activity patterns that are, at once, distributed within a system of interacting cortical areas (integrated) and differentiated in space and time (information rich). We thus hypothesized that the level of consciousness could be estimated empirically by perturbing the cortex with TMS to engage distributed interactions and by measuring the information content (algorithmic complexity) of the resulting response, as measured with EEG. We found that the algorithmic complexity of the spatial-temporal pattern of electro-cortical activation triggered by transcranial magnetic stimulation reliably discriminated the level of consciousness in single individuals across different conditions - including wakefulness, dreaming, the locked-in syndrome, the minimally conscious state, the vegetative state/unresponsive wakefulness syndrome, NREM sleep, midazolam deep sedation, xenon and propofol anesthesia. These empirical results support the theoretical notion that consciousness is linked to the complexity of distributed interactions within the corticothalamic system and provide a principled approach for estimating the level of consciousness at the bedside.

Selected references:
Rosanova M, Gosseries O, Casarotto S, Boly M, Casali AG, Bruno MA, Mariotti M, Boveroux P, Tononi G, Laureys S, Massimini M (2012). Recovery of cortical effective connectivity and recovery of consciousness in vegetative patients. BRAIN, vol. 135, p. 1308-1320

M. Massimini, M. Boly, A. Casali, M. Rosanova, G. Tononi (2009). A perturbational approach for evaluating the brain's capacity for consciousness. PROGRESS IN BRAIN RESEARCH, vol. 177, p. 201-211

M. Massimini, F. Ferrarelli, R. Huber, S.K. Esser, H. Singh, G. Tononi (2005). Breakdown of cortical effective connectivity during sleep. SCIENCE, vol. 309, p. 2228-2232

Date: November 21, 2012
Speaker: Tomi Peltola (BECS)
Topic: Genome-wide association analysis and Bayesian variable selection.

Abstract: The progress in high-throughput measurement technologies has enabled cost-effective measurement of the genetic make-up of individuals at the scale of millions of markers along the genome. During the last decade or so, genome-wide association studies have led to numerous findings linking genetic variation to disease and other traits.

The primary statistical analysis is often conducted with simple methods to ease the computation. However, the methods ignore basic assumptions about the nature of the data, such as that multiple genetic markers are thought to contribute to many diseases and traits. An alternative approach is provided by Bayesian variable selection, where the full set of measured genetic markers is modelled simultaneously.

The presentation will include a description of the genetic data used in genome-wide association studies, a brief view to the primary analysis and a description of the Bayesian approach with its potential benefits and challenges.

Date: November 14, 2012
Speaker: Olli-Pekka Mutanen (Aalto Ventures Program)
Topic: Value co-creation models in teaching and case entrepreneurship education in Aalto.

Abstract: Entrepreneurship and growth firms have never had more societal importance as the source of wealth and new job creation. For example, over the last few decades, nearly all the economic growth and increase in employment in the U.S. has come from high-growth technology companies (Startup Genome, 2012). It seems to be evident that new jobs will be born mostly in young but rapidly scaling firms in the future. But how does our university education take this remarkable change into account?

Aalto University’s answer to this challenge and opportunity is at least twofold: First, promoting a culture of creativity and entrepreneurship is a central goal both in teaching and societal level impact and part of Aalto's strategy (source: Strategic Development of Aalto University, 2012). Secondly, Aalto Ventures Program, a master´s level minor program in growth entrepreneurship was launched in the beginning of the current academic year. The program aims to deliver students entrepreneurial attitude, knowledge & skills, and make them work in cross-disciplinary teams with real life business cases supported by mentoring from experienced serial entrepreneurs and investors. In addition to this particular program, it is clear that university level goals regarding entrepreneurship implicate also other requirements and changes in our teaching methods and curricula in overall. The big question is what kind of changes and how they should be implemented?

This talk tries to give some preliminary answers by dealing with the importance of entrepreneurial thinking, creation of innovation based growth firms and their implications to our educational goals and teaching methods. We also will have a look at the Aalto Ventures Program and demonstrate one of its courses planned to create value both to students and startup entrepreneurs. The aim of the talk is to promote entrepreneurial thinking in the context of teaching and perhaps to give also some ideas to plan win-win settings in our future teaching.

Date: November 7, 2012
Speaker: Vasyl Palchykov (BECS)
Topic: Irregularity of communication patterns: a study of mobile phone networks.

Abstract: Many real systems are characterized by high level of spatial and temporal inhomogeneities. Mobile phone communication networks are among them: besides the geographical diversities, individuals display high level of temporal irregularity in their communication patterns. In this short seminar I will talk about the ways how to measure these irregularities and what we can learn on their basis.

Date: October 31, 2012 (canceled due to illness)
Speaker: Tomi Peltola (BECS)
Topic: Genome-wide association analysis and Bayesian variable selection.

Abstract: The progress in high-throughput measurement technologies has enabled cost-effective measurement of the genetic make-up of individuals at the scale of millions of markers along the genome. During the last decade or so, genome-wide association studies have led to numerous findings linking genetic variation to disease and other traits.

The primary statistical analysis is often conducted with simple methods to ease the computation. However, the methods ignore basic assumptions about the nature of the data, such as that multiple genetic markers are thought to contribute to many diseases and traits. An alternative approach is provided by Bayesian variable selection, where the full set of measured genetic markers is modelled simultaneously.

The presentation will include a description of the genetic data used in genome-wide association studies, a brief view to the primary analysis and a description of the Bayesian approach with its potential benefits and challenges.

Date: October 24, 2012
Speaker: Roman Volinsky (BECS)
Topic: Introduction to Nano-Jewelry.

Abstract: Construction of controlled, organized nanostructure assemblies has been among the most challenging aspects of nanotechnology research and development. Diverse approaches for the design and fabrication of nanostructures on solid surfaces have been reported. While "top-down" lithography techniques have predominated, new "bottom-up" self-assembly methods have garnered increasing interest as viable alternatives. In particular, self-assembled monolayers at the air/water interface (i.e. Langmuir monolayers) have been proposed as promising vehicles for surface patterning of metallic and semiconductor nanostructures.

Here we demonstrate that lipid monolayers that are isothermally compressed at the air/water interface can serve as templates for novel surface architectures of gold nanostructures. Specifically, we show that distinct patterns of alkylthiol-capped Au NPs can be formed within the diacetylene films. The surface patterns can be transferred from the water surface onto solid substrates and further annealed to remove the organic template, yielding organized Au patterns and pointing to the potential utilization of this approach for fabrication of surface nanostructures. In addition, we describe unique long-range ordering of gold nanoparticles that occurs when mixed monolayers of the NPs and an unsaturated fatty acid are irradiated with low-intensity laser light. We demonstrate the formation of a network of elongated Au "wires" occurred only at zero surface pressure, in which the mixed monolayers exist in highly fluid phases. We show that this unusual and seemingly counterintuitive phenomenon is reversible and is due to local heating of the monolayer surface by the laser irradiation in a very narrow temperature range.

Date: October 17, 2012
Speaker: Tuomas Hirvonen
Topic: Relationship between electrophysiological and hemodynamic markers of neural activity in cognitive neuroimaging

Abstract: Both magnetoencephalography (MEG) and functional magnetic resonance imaging (fMRI) provide an opportunity to study human brain function nonivasively. MEG measures directly the electrical activity of the cells in the cerebral cortex. FMRI, instead, is an indirect method: it measures changes in cerebral blood flow and hemoglobin concentration which are believed to be driven by the increased energy demand of active neural populations. However, the relationship between fMRI signals and electric activity of nerve cells, commonly referred to as neurovascular coupling, is far from simple.

The purpose of this study was to examine spatiospectral heterogeneity of correlations between fMRI and MEG measurements; in particular, the goal was to investigate neurovascular coupling patterns during cognitive tasks. We wanted to identify regions that show significant correlation between electric and hemodynamic signals, and to examine what kind of correlation patterns can be found between the fMRI and the frequency-decomposed MEG signals in these regions. As previous studies have mostly used simple tasks and concentrated on primary sensory cortices, a central goal was to find out if the previously detected patterns also hold for other brain areas and in more complicated cognitive tasks. The analysis was conducted with the Partial Least Squares Correlation (PLSC) method.

Date: October 10, 2012
Speaker: Andrey Zhdanov (BECS)
Topic: Capturing behavioral aspects of functional brain imaging experiments

Abstract: Developers of functional brain imaging instrumentation, whether EEG, MEG, fMRI or anything else, focus primarily on recording subject's brain signals. Behavioral aspects of the experiment attract relatively little attention and are typically left over to third-party stimulus presentation systems. This approach essentially attempts to "force" the required behavioral state on the subject in a controlled fashion.

An alternative is to record relevant aspects of subject's behavior together with brain signals. This approach treats recording of behavioral data as an integral part of the imaging procedure rather than just an add-on. In this talk I will describe several related projects in which behavioral audio/video recordings were integrated into TMS and MEG experiments and share the lessons learned.

Date: October 3, 2012
Speaker: Hang-Hyun Jo (BECS)
Topic: Spatiotemporal correlations of handset-based service usages

Abstract: We study spatiotemporal correlations and temporal diversities of handset-based service usages by analyzing a dataset that includes detailed information about locations and service usages of 124 users over 16 months. By constructing the spatiotemporal trajectories of the users we detect several meaningful places or contexts for each one of them and show how the context affects the service usage patterns. We find that temporal patterns of service usages are bound to the typical weekly cycles of humans, yet they show maximal activities at different times. We first discuss their temporal correlations and then investigate the time-ordering behavior of communication services like calls being followed by the non-communication services like applications. We also find that the behavioral overlap network based on the clustering of temporal patterns is comparable to the communication network of users. Our approach provides a useful framework for handset-based data analysis and helps us to understand the complexities of information and communications technology enabled human behavior.

Date: September 26, 2012
Speaker: TEDx Oxbridge online talk by David Spiegelhalter with introduction and Q&A by Aki Vehtari (BECS)
Topic: Motorbikes, Terrorism, Heart Attacks, Sausages

Abstract: David Spiegelhalter's background is in medical statistics, particularly the use of Bayesian methods in clinical trials, health technology assessment and drug safety. In his post as Winton Professor for the Public Understanding of Risk he leads a small team ( that attempts to improve the way in which the quantitative aspects of risk and uncertainty are discussed in society. He gives many presentations to schools and others, advises organisations on risk communication, and is a regular newspaper columnist on current risk issues. He has also appeared on Winter Wipeout. He was elected FRS in 2005 and awarded an OBE in 2006 for services to medical statistics.

Date: September 18, 2012
Speaker: Suvi Talja
Topic: Temporal Dynamics of Task-Dependent Activations of Human Auditory Cortex: An EEG Study

Abstract: Rinne at al. (2009) reported that human auditory cortex activations measured with fMRI are strongly dependent on the auditory task. Although fMRI has a high spatial resolution its temporal accuracy is not sufficient to examine the activation order of different areas of the auditory cortex during active listening tasks. The present study tested whether these task-dependent activations can be investigated using source estimation of scalp-recorded auditory evoked potentials.

Subjects (17) performed pitch discrimination, pitch memory (three difficulty levels in both tasks) or visual tasks during the presentation of sounds. The auditory evoked potentials were recorded using 136 scalp electrodes. Sources of the evoked potentials were modeled using cortically constrained and depth- and orientation-weighted minimum norm estimation.

The evoked potentials were modulated by task at 200-700 ms from sound onset. Source estimation revealed stronger activation (350-700 ms) in the anterior auditory cortex of the left hemisphere during pitch discrimination than during visual task with the same sounds. Pitch memory task, in turn, was associated enhanced activation (vs. visual task) in the bilateral inferior parietal lobule (500-650 ms) and decreased activation (vs. visual and discrimination tasks) in the left anterior superior temporal gyrus (200-300 ms). These task-dependent activations are very similar to those reported in the previous fMRI study. Therefore, the results of the thesis indicate that evoked potentials can be used to obtain temporal information on the task-dependent activations of human auditory cortex.

Rinne T, Koistinen S, Salonen O & Alho K (2009). Task-Dependent Activations of Human Auditory Cortex during Pitch Discrimination and Pitch Memory Tasks. J Neurosci 29:13338-13343.

Seminar schedule, spring 2012

Date: June 13, 2012
Speaker: Dr. Alex Petersen (IMT Lucca, Italy)
Topic: A microscopic perspective on academic career growth: empirical analysis and theoretical models

Abstract: The age of "micro sociology" has arrived, brought forth by the immense quantity of disambiguated data which is providing new grounds for understanding human dynamics at the individual scale. In this talk I will present a data-centric picture of academia focusing on the distinctive features of knowledge and collaboration spillovers, cumulative advantage, and competition.

A central motivation is to better understand how institutional changes within academia (i.e. shifts away from tenure towards shorter-term contracts) may affect the overall potential of science and scientists. We use quantitative patterns in the longitudinal publication data of 300 leading physicists and cell biologists and 100 assistant professors in physics to develop and test models for career growth in competitive systems. Our theoretical model shows that short-term contracts can amplify the effects of competition and uncertainty making careers more vulnerable to early termination, not necessarily due to lack of individual talent and persistence, but because of random negative production shocks. We find corroborating evidence, using the publication data of ~400,000 scientists taken from six high-impact journals (Nature, Science, PNAS, CELL, NEJM, and Phys. Rev. Lett.), which further highlights the importance of early career development by showing that a significant fraction of careers can be stunted by the relative disadvantage associated with inexperience.

Date: June 6, 2012, at 1:15 PM (coffee served at 1:00 PM)
Speaker: Dr. Tapani Salmi (University of Helsinki)
Topic: Mobile Phone Technology in Diagnostics and Screening of Sleep Disorders

Date: May 30, 2012
Speaker: Marija Mitrović (BECS)
Topic: Structure and dynamics of communities on blogs

Abstract: The Internet experience in recent years has revolutionized the mechanics that individual can exploit to participate in global social dynamics. Indirect on-line communication at various Web portals leads to massive data and potentially new techno-social phenomena. We present the quantitative method for studying of structure and dynamics of cybercommunities, which combines the approaches of theory of complex networks and physics of complex systems. With the machine learning methods we classify the texts of posts and comments for their emotional contents. The data obtained from blogs and similar websites are mapped onto bipartite networks of users and their comments written on posts. The topological properties of these networks and their projections are closely related to blogging habits of users. Specifically, clusters of users, topological communities, detected in weighted projections evidence the existence of two different types of blogging dynamics regarding post popularity. The statistical analysis of temporal activity of users, avalanches of emotional comments and power spectrum of time series, indicate the system which exhibits self-organized critical behavior and temporal correlations.

Date: May 23, 2012
Speaker: Dr. Markus Johansson (Chalmers University of Technology, Gothenburg, Sweden)
Topic: Modeling of Electromagnetic Sources using Phase Retrieval Methods

Abstract: Modeling of the field distribution from electromagnetic sources is of interest for various applications including electromagnetic compatibility investigations and electromagnetic dosimetry. Gradient based optimization algorithms for determining the total field, including phase information, when only field amplitudes have been measured in front of an electromagnetic source have been developed. Results for the phase retrieval methods for numerical test cases as well as measured field will be presented.

Date: May 2, 2012
Speaker: Stina Immonen (Aalto SCI)
Topic: On the ongoing study reform of the Aalto Bachelor's degree studies

Stina Immonen from Aalto SCI will visit BECS and give a 20-30 minute presentation on the Bachelor studies' reform that is taking place in Aalto university. She will also tell us the latest news on the plans regarding Master's degree (DI) studies. After the presentation it is possible to continue with some discussion and questions. So everyone, this is a good opportunity for getting information regarding these important changes in teaching at Aalto and SCI School.

Date: April 27, 2012
Speaker: Will Cardwell, Head of Aalto Centre for Entrepreneurship (ACE), and Zaira Mammadova, Communications Coordinator at ACE
Title: Introduction to ACE and AppCampus

Abstract: ACE: Aalto Center for Entrepreneurship (ACE) aims at creating business success stories from the science and art within Aalto community, and working as a catalyst for elevating high ambition entrepreneurship from Finland and through the Baltic region.

AppCampus: To drive innovation and business opportunities in Finland's mobile ecosystem and beyond, Microsoft Corp. and Nokia will each invest up to 9 million euros into a newly established mobile application development program at Aalto University during the next three years. The AppCampus program has been set up to foster the creation of innovative mobile applications for the Windows Phone ecosystem, and in addition, Nokia platforms, including Symbian and Series 40, to create a new generation of self-sustaining mobile startups.

Date: April 18, 2012
Speaker: Arto Meriläinen
Title: Adapting a Gaze Tracking System to a Mobile Environment

Abstract: Gaze tracking has traditionally been performed in controlled laboratory environments. The experiment setups have commonly been limited to study only computer-human interaction. Recently, the need to perform experiments in natural environments has emerged in different areas of science.

This thesis overviews the structure of a mobile gaze tracking system that was developed in the Ganzheit project. The goal of the gaze tracking system is to offer an open alternative for commercial systems. The developed gaze tracker utilises model based gaze tracking. The approach is accurate and robust against movements of the head.

In order to utilise model based gaze tracking, it is vital to identify the pupil and corneal reflections from an eye image. The constructive part of the thesis focuses on developing a method for identifying the features from the eye image. The developed method is tested with experimental data. The results show that the method for finding the pupil is accurate and robust against differences in facial features and changing lighting conditions. The goodness of recognising the corneal reflections varies between test subjects.

The mobile gaze tracking system is experimented in an ordinary office room. The results indicate that the developed method works adequately.

Date: April 11, 2012
Speaker: Arnab Chatterjee (BECS)
Title: Of social efficiency and resource allocation

Abstract: We discuss a class of resources allocation processes, which are relevant to a variety of disciplines including computer science, biology, social sciences. We study the statics and dynamics of such problems in terms of social efficiency while the agents have zero to minimal coordination, and also study whether the state of maximal efficiency is eventually attained when some minimum global information is provided.

Date: April 4, 2012
Speaker: Lauri Parkkonen (BECS)
Title: Using MEG to look at brain states rather than transient brain activations

Abstract: Non-invasive time-resolved brain imaging tools such as electroencephalography (EEG) and magneto­encephalography (MEG) excel in picking up transient brain responses to changing sensory input but may not perform as well in differentiating endogenous “brain states” such as the target of attention or the percept when viewing an ambiguous scene. This problem can be partially addressed by using temporally structured, or tagged, stimuli as probes to indirectly measure these states. Tagging can be applied to basically any sensory modality but most commonly it is used for audition and vision by modifying the original stimulus either by amplitude modulation or by superimposing an additional signal. We have employed tagging and MEG to study visual perception, the effects of selective attention and their utility in providing real-time feedback to the subject. I will also discuss our preliminary results on dichotic listening to tagged auditory streams and on attentional selection of multiple tagged regions in the visual field.


Fujiki N, Jousmäki V & Hari R (2002): Neuromagnetic responses to frequency-tagged sounds: a new method to follow inputs from each ear to the human auditory cortex during binaural hearing. J Neurosci 22:RC205.

Parkkonen L, Andersson J, Hämäläinen M & Hari R (2008): Early visual brain areas reflect the percept of an ambiguous scene. Proc Natl Acad Sci U S A 105:20500-4.

Sudre G, Parkkonen L, Bock E, Baillet S, Wang W & Weber DJ (2011): rtMEG: a real-time software interface for magnetoencephalography. Comput Intell Neurosci 2011:327953.

Date: March 21, 2012
Speaker: Andrey Zhdanov (BECS)
Title: Combined EEG-fMRI recordings: overview and a bit of clinical perspective

Abstract: Electroencephalography (EEG) and functional Magnetic Resonance Imaging (fMRI) are two very different methods of studying brain activity. The information that they provide is in many respects complementary, however recording them simultaneously is technically challenging. While combined EEG-fMRI recordings are becoming an established research (and to a lesser degree also clinical) tool, the technical problems are still far from being solved. In the talk I will give a short introduction to combined EEG-fMRI recordings, describe the technical challenges and the way they are addressed and discuss their applications, in particular, in treatment of epilepsy.

Date: March 14, 2012
Speaker: Arno Solin (BECS)
Title: Filtering of Physiological Noise in BOLD fMRI Using DRIFTER

Abstract: The DRIFTER algorithm [1] is a new model-based Bayesian method for retrospective elimination of physiological noise from functional magnetic resonance imaging (fMRI) data. In the method, we first estimate frequency trajectories of the physiological signals with the interacting multiple models (IMM) algorithm. The frequency trajectories can be estimated from external reference signals, or if the temporal resolution is high enough, from the fMRI data. The estimated frequency trajectories are then used in a state space model in combination of a Kalman filter (KF) and Rauch-Tung-Striebel (RTS) smoother, which separates the signal into an activation related cleaned signal, physiological noise, and white measurement noise components. The components can be subject to further study [2] or then simply used to eliminate the periodic noise in the data in order to achieve higher signal-to-noise ratios.

[1] Särkkä S., Solin A., Nummenmaa A., Vehtari A., Auranen T., Vanni S., Lin F.-H. (2012). Dynamic Retrospective Filtering of Physiological Noise in BOLD fMRI: DRIFTER. NeuroImage, 60:1517-1527.
[2] Särkkä S., Solin A., Nummenmaa A., Vehtari A., Auranen T., Vanni S., Lin F.-H. (2012). Identification of Spatio-Temporal Oscillatory Signal Structure in Cerebral Hemodynamics Using DRIFTER. To appear as abstract and e-poster in ISMRM 2012.

Date: March 7, 2012
Speaker: Enrico Glerean (BECS)
Title: FMRI phase synchronization as a measure of dynamic functional connectivity

Abstract: Brain functional connectivity and activation have been studied with methods such as seed-based and intersubject correlation of functional magnetic resonance imaging (fMRI) data. To inspect brain temporal dynamics using such correlation-based methods, sliding time windows have been used with necessary restrictions on the size of the temporal window to avoid estimation bias of the correlation coefficients. We hypothesized that it is possible to increase the temporal resolution of functional connectivity analysis by using instantaneous phase synchronization (PS) as a measure of dynamic (time-varying) functional connectivity. Here, we applied PS on an fMRI dataset obtained while 12 healthy volunteers watched a feature film. We used a narrow frequency band (0.04 – 0.07 Hz) in the PS analysis to avoid artefactual results. We defined three metrics for computing time-varying functional connectivity and time-varying intersubject reliability based on estimation of instantaneous PS: i) seedbased phase synchronization (SBPS), ii) intersubject phase synchronization (IPS), and iii) intersubject seed-based phase synchronization (ISBPS). Our findings show that these PS-based metrics yield results consistent with both seed-based correlation and intersubject correlation methods, while providing maximal temporal resolution. These metrics can be applied both in studies with complex naturalistic stimuli (e.g., watching a movie or listening to music in the MRI scanner) and more controlled paradigms (e.g., event related or blocked design) as long as identical stimulus sequence is presented to all subjects. A Matlab toolbox – FUNPSY – is released for using these metrics in fMRI data analysis.

Date: February 29, 2012
Speaker: Ajay Mahalka (BECS)
Title: Protein-Phospholipid Interactions: From Biophysics to Therapeutics

Abstract: Lipid–protein interactions play a key role in a large number of cellular processes and are controlled by the membrane associated physicochemical properties and structure of membrane lipids. Recently, several lines of evidence have merged to suggest that membranes induced cytotoxic oligomers formation are responsible for killing of cells and loss of tissue function in Alzheimer's and Parkinson's disease, prion disease, and type 2 diabetes [1,2]. Interestingly, the same underlying mechanism involving lipids has been recently concluded to be responsible also for the targeting of host defence proteins [3]. We also demonstrated that oxidized phospholipid membrane accelerates Finnish type familial gelsolin amyloidosis in vitro [3]. In addition to counteracting protein aggregation, the evolutionarily conserved chaperone Hsp70 (heat shock protein 70) promotes cell survival by, e.g., inhibiting the permeabilization of lysosomal membranes [5,6]. In conclusion protein-phospholipid interactions open novel venues for the development of therapeutics for amyloid, lysosomal storage disorders (Niemann-pick disease), and cancer.

1. Kinnunen PKJ, Amyloid formation on lipid membrane surfaces, The Open Biology Journal 2 (2009) 163-175
2. Kinnunen PKJ, Mahalka AK: Protein-Oxidized Phospholipid Interactions in Cellular Signalling for Cell Death: From Biophysics to Clinical Correlations (submitted to BBA, Biomembrane)
3. Mahalka AK, Kinnunen PKJ: Binding of amphipathic alpha-helical antimicrobial peptides to lipid membranes: lessons from temporins B and L, BBA, Biomembrane 2009, 1788, 1600-1609.
4. Mahalka AK, Maury CPJ, Kinnunen PKJ: 1-palmitoyl-2-(9'-oxononanoyl)-sn-glycero-3-phosphocholine, an oxidized phospholipid accelerates Finnish type familial gelsolin amyloidosis in vitro, Biochemistry 2011, 50, 4877–4889
5. Kirkegaard T, Roth AG, Petersen NH, Mahalka AK, Olsen OD, Moilanen I, Zylicz A, Knudsen J, Sandhoff K, Arenz C, Kinnunen PKJ, Nylandsted J, Jaattela M: Hsp70 stabilizes lysosomes and reverts Niemann-Pick disease-associated lysosomal pathology, Nature 2010, 463, 549-553.
6. Mahalka AK, Code C, Kirkegaard T, Jaattela M, Kinnunen PKJ: Activation of phospholipase A2 by Hsp70 in vitro, BBA Biomembrane 2011,1808, 2569-2572

Date: February 15, 2012
Speaker: Hang-Hyun Jo (BECS)
Title: The origin of bursts in human communication: from weekly cycles to risk-averse behavior

Abstract: The temporal communication patterns of human individuals are known to be inhomogeneous or bursty, which is reflected as the heavy tails in the inter-event time and waiting time distributions. As the cause of such bursty behavior two main mechanisms have been suggested: a) Inhomogeneities due to the circadian and weekly activity patterns and b) inhomogeneities rooted in human task execution. We investigate the roles of these mechanisms by developing and then applying systematic de-seasoning methods to remove the circadian and weekly patterns from the time series of mobile phone communication events of individuals. We find that the heavy tails in the inter-event time distributions remain robust with respect to this procedure, which clearly indicates that the detailed understanding of the role of human factors in bursty dynamics is required. In order to investigate their role we devise a risk-averse agent model, where an agent in an uncertain situation tries to reduce the uncertainty by communicating with information providers while having to wait time for responses. Here the waiting time can be considered as cost. We show that the optimal choice of the waiting time under uncertainty gives rise to the bursty dynamics, characterized by the heavy tailed distribution of optimal waiting time. We find that in all cases the efficiency for communication is relevant to the scaling behavior of the optimal waiting time distribution. On the other hand the cost turns out in some cases to be irrelevant depending on the degree of uncertainty and efficiency.

Date: February 8, 2012
Speaker: Isambi Mbalawata
Title: Parameter Estimation in Stochastic Differential Equations with MCMC and Kalman Filtering

Abstract: The presentation is about parameter estimation in stochastic differential equations (SDE) using Markov chain Monte Carlo (MCMC) methods. The MCMC methods used are Metropolis–Hastings and Hamiltonian Monte Carlo (HMC). We explain how Kalman filter can be used to compute the marginal likelihood and the energy function needed in the MCMC methods. In the case of SDEs, computation of the energy function gradient needed by HMC and gradient based optimization methods is non-trivial, and here we explain how the gradient can be efficiently computed with a Kalman filter-like recursion. To demonstrate the method we use the Biochemical Oxygen Demand model.

Date: February 1, 2012
Speaker: Santo Fortunato (BECS)
Title: Statistical physics of citations

Abstract: Citation behavior has been subject of intense investigations over the last years. The availability of detailed databases and of modern computers enables one to perform careful statistical analyses of citation data and their patterns. One of the results of these investigations is the fact that pure citation scores are not reliable to provide fair rankings between papers and/or scientists, for several reasons. One of these reasons is the role played by the specific scientific discipline of a paper/author. Here I show that the citation patterns of papers of different disciplines are actually identical, provided the citation scores are properly normalized. This provides a criterion for an objective comparison of papers and scientists belonging to different disciplines. Another improvement may come from a self-consistent weighing of citations, based on the role of scientists in the spreading of reputation to their peers, alike to Google’s PageRank process.

Seminar schedule, fall 2011

Date: December 21, 2011
Speaker: Koen Van Leemput (BECS)
Title: Computational analysis of structural brain MR - towards a useful tool in clinical practice

Abstract: Computational methods that can reliably delineate structures of interest from magnetic resonance (MR) images of the brain have a wide array of potential applications, including assisting in clinical diagnosis and therapy monitoring, as well as facilitating basic neuroscientific and clinical research. In order to reach their full potential in the clinic, such methods will need to be complemented by multi-variate regression and classification models that can predict future clinical outcomes based on automatically obtained measurements of brain integrity.

In this talk I will discuss some of the technical difficulties associated with solving these tasks, and highlight our ongoing research based on generative and discriminate Bayesian models of brain MR.

Date: November 30, 2011
Speaker: Jaakko Kauramäki (BECS)
Title: Top-down mediated enhancements and suppressions in the human auditory cortex

Abstract: In this talk I will present four EEG/MEG experiments to be included in my PhD thesis. These explore the neural basis of auditory cortex top-down modulations due to 1) selective attention and 2) cross-modal interactions while lipreading. The results are opposite at the evoked-response level: selective attention enhances and lipreading suppresses the N100/N100m and later-latency responses, but both effects are in some cases feature-selective as we shall see from the results.

Date: November 23, 2011
Speaker: Margareta Segerståhl (in place of Santo Fortunato), BECS
Title: Teaching and graduate studies at BECS, coordinator's perspective

Abstract: In this presentation (about 20min) I first tell about my coordinator role at BECS. Then I give some info for all those who are formally in change of courses that BECS provides and for all of you who are involved in the actual teaching! Another important issue for all our graduate students and their supervisors is the overall role of graduate students in BECS teaching and everyday functioning. I hope that the audience also participates in this presentation and we could hear questions, comments and feedback that might be useful for many of us.

Date: November 16, 2011 (canceled due to overlapping installation lectures)
Speaker: Andrey Zhdanov (BECS)
Title: Combined EEG-fMRI recordings: overview and a bit of clinical perspective

Abstract: Electroencephalography (EEG) and functional Magnetic Resonance Imaging (fMRI) are two very different methods of studying brain activity. The information that they provide is in many respects complementary, however recording them simultaneously is technically challenging. While combined EEG-fMRI recordings are becoming an established research (and to a lesser degree also clinical) tool, the technical problems are still far from being solved. In the talk I will give a short introduction to combined EEG-fMRI recordings, describe the technical challenges and the way they are addressed and discuss their applications, in particular, in treatment of epilepsy.

Date: November 9, 2011
Speaker: Sanjeev Ranjan (HBBG, BECS)
Title: Liposome nanoparticles for targeted drug delivery, gene delivery, and magnetic imaging

Abstract: Investigations aimed at developing multifunctional lipid-based nanoparticles (LNPs) to deliver drug and gene with peptide driven targetibility and MRI visualization using the inner ear as a model target organ for developing therapies of inner ear disorders. The LNPs are targetable to selected cell populations, biodegradable, traceable in-vivo, and equipped with controlled drug release. The diameter of the nanoparticles is an important attribute to enable them to overcome the various in vivo barriers for systemic delivery such as: the blood components, reticuloendothelial system uptake, tumor access, extracellular matrix components, and intracellular barriers. As a 1st aim, the main effort has been given to prepare LNPs with a diameter less than 100 nm using a novel procedure, adaptive focused ultrasound (AFU) [1]. AFU has several advantages compared to other techniques as it is non-invasive, isothermal and the energy involved is much more precisely controlled due to focusing of the acoustic energy. The LNPs formed are free of organic solvents and detergents. The development of new methods for the efficient delivery of drugs into the inner ear represents an important step towards the treatment of cochlear diseases or injury and the amelioration of hearing loss. 2nd aim of this project was to analyze the efficacy of the penetration of LNPs through the round window membrane when injected into the middle ear cavity in mice and to determine whether LNPs accumulating in the inner ear tissues was sufficient to elicit a therapeutic response. Our results demonstrate that LNPs are capable of carrying a drug (Disulfiram) into the inner ear that elicits a biological effect, with consequences measurable by a functional readout [2]. Under 3rd aim the study investigated for selective targetability of inner ear neuronal cell survival receptor tyrosine kinase B (TrkB) for targeted gene delivery. We demonstrate the feasibility of targeting of LNPs to TrkB expressing cells by designed peptides, promoting cellular uptake via receptor-mediated pathways [3, 4]. Our 4th aim was to track the dynamics and distribution of LNPs in vivo by preparing MRI traceable LNPs and their distribution in the inner ear. Effective MRI traceable LNPs were developed by encapsulating Gadolinium which were visualized in vivo in the rat inner ears [5]. Presently we are studying the impact of size on transportation efficacy of LNPs across the middle-inner ear barriers and the trafficking mechanism of these LNPs in auditory nerve cells. We are also studying the biophysical characteristics of our designed pH sensitive endosomolytic peptides which have shown enhanced gene expression in mammalian cells. A patent filing is in process at Aalto University claiming novel designing of endosomolytic peptides.

1. Langmuir, 2011, 27 (16), pp. 10088-10097
2. Nanomedicine, 2011, submitted
3. The Journal of Gene Medicine, 2011, 13 (2), pp. 134-144
4. European Journal of Nanomedicine, 2009, 2 (2), pp. 7-13
5. Journal of Nanobiotechnology, 2010, 8: 32

Date: November 2, 2011
Speaker: Raj Kumar Pan (BECS)
Title: A task execution model

Abstract: In this talk I will express my views about why physicist are obsessed with models. In keeping with the tradition, I present a new model to explain how humans execute different tasks. In this model an agent tries to execute a task, however, he often get disrupted by other factors. As with all good models, I show that our model is in line with some of the empirically observed dataset. I will end the talk with a grand conclusion, claiming that our analysis helps to understand the human task execution behavior in a realistic scenario.

Date: October 26, 2011
Speaker: Krisztina Cziner (Aalto Research Support Services)
Topics: Research and Development funding strategies including EC policies; National R&D Funding Possibilities, How to write a competitive proposal (Presentation in PPTX format)

Date: October 19, 2011
Speaker: Janne Ojanen (BECS)
Title: Bayesian predictive model selection

Abstract: Model selection is one of the important open problems in statistics. While a standard scientific approach is to compare theories based on the accuracy of the predictions they give, in (Bayesian) statistics there is a multitude of different model selection approaches all claiming to be predictive.

The talk is a gentle introduction into the basic concepts of Bayesian predictive model selection. The aim is to give the audience an understanding on how the assumptions behind the formulation of model selection task as a statistical decision problem affect the model selection results.

Date: October 12, 2011
Speaker: Santeri Yrttiaho (BECS)
Title: Cortical processing of the periodicity of speech sounds

Abstract: The periodicity of speech sounds plays a significant role in speech communication. The studies presented in the talk investigated cortical processing of the periodicity of speech sounds with magnetoencephalography by measuring brain activation elicited by vowel stimuli with variable periodic structures. The results indicate larger amplitudes and more anterior source locations for the responses elicited by periodic as opposed to aperiodic vowel stimuli. While such an effect of periodicity was observed for a range of stimulus parameters, cortical periodicity-specific activation was also modulated by the features of vowel stimuli. Furthermore, evidence for aperiodicity-sensitive activation was found through stimulus-specific release from adaptation when aperiodic vowel stimuli were alternated with periodic rather than with aperiodic adaptors.

The results of the studies, thus, indicate that the degree of speech sound periodicity is represented in the auditory cortex. Such sensitivity to periodicity might reflect the activation of distinct neural populations that are selective to sound periodicity and aperiodicity. Importantly, this view of distinct feature-selective populations could now be generalized to describe cortical processing of speech. The dependency of the observed periodicity-sensitivity on the acoustic features of the vowel stimuli, further, appears to reflect cortical encoding of auditory-perceptual aspects of voice quality.

Date: September 28, 2011
Speaker: Lauri Kovanen (BECS)
Title: Temporal motifs

Abstract: Large empirical social networks are usually obtained by aggregating temporal data; for example the mobile phone networks are created by aggregating mobile phone calls over some time period. This aggregation discards all information about the timing of the events, something that has turned out to be highly non-trivial. We define temporal subgraphs so that they include information not only about the topology but also about the mutual order of the events, and then divide the temporal subgraphs into equivalence classes called temporal motifs. Our implementation allows identifying temporal motifs in a mobile phone data set with hundreds of millions of calls, and the results highlight well the role that burstiness and causality play in communication. The method also allows studying differences in temporal motifs between different users groups, defined for example by gender and age, as well as between different times of the day.

Date: September 14, 2011
Speaker: Mikko Viinikainen (BECS)
Title: Parametric approach on the neural correlates of valence

Abstract: An important framework in looking into emotions is emotion dimensions such as valence, i.e. the unpleasantness – pleasantness dimension. The prevailing view that valence is a single dimension has been challenged by studies, which show that negative valence and positive valence can be in some cases independent. Functional magnetic resonance imaging (fMRI) data supporting separate negativity and positivity representations in the brain is presented and discussed.

Date: September 7, 2011
Speaker: Kirsi Kettula (Educational Developer, Strategic Support for Research and Education / Educational Development Team)
Title: Conference presentation – A daydream or a nightmare?

Abstract: Standing in front of an audience is not an easy task. But neither are swimming, riding a bicycle – or doing research! Yet, we have learned many skills that seemed difficult at the beginning. To give a conference presentation is just a skill among other skills, and thus it can be learned. No one is born as a keynote speaker! There are several hints and tips that can help you to start with, but after that it is only about practising, practising, practising - and some constructive feedback.

Seminar schedule, spring 2011

Date: June 22, 2011
Speaker: Julio C Hernández Pavón (TMS—EEG, BECS)
Title: General aspects of TMS—EEG

Abstract: The combination of transcranial magnetic stimulation (TMS) and electroencephalography (EEG) provides us the possibility to non-invasively evaluate cortical excitability and functional connections between different areas of the brain. In this talk, I will describe the physical and biological principles of TMS—EEG as well as potential applications to study functional connectivity, brain mapping and neuroplasticity.

Date: June 8, 2011
Speaker: Hanna Heikkinen (Biophysics, BECS)
Title: How reduced models in electrophysiological research correspond to system function: a case study on rod photoreceptors

Abstract: Experimental studies on neuronal networks may range from recording field potentials from live subject to single cell electrophysiology and even biochemical assays. Reduction and invasiveness brings resolution to understanding the details of a network, yet it gets more difficult to relate the results to its natural function. Retina, the “Nature’s brain slice”, is a neuronal network with well characterized input and relatively well understood architecture and circuitry. Photoreceptors, which provide the first step of retinal signal processing, are well understood models of transduction, regulation and transmission of cell level signals. However even in the rod photoreceptor, whose function is understood with amazing detail, it is not clear how well various experimental data reflects the natural function of the cell. We compared the electrical (ERG) photoresponses of mouse rods in anesthetized animals and in the intact isolated retina. The gain and kinetics of the photoreceptors’ transduction process were well preserved upon isolation, but easily disturbed by manipulating the experimental conditions. On the other hand, the function of the rest of the retina was altered from the very first synapse onwards, emphasizing the importance of assaying the physiological state of neural tissues in electrophysiological studies.

Date: May 25, 2011
Speaker: Brannon Green (Georgetown University)
Title: The neural substrates of audition in humans and non-human primates

Abstract: The Laboratory of Integrated Neuroscience and Cognition (LINC) investigates the functional organization and plasticity of the cerebral cortex. Research follows three general themes: processing streams in audition (space and objects, music and speech); brain reorganization in tinnitus; and cross-modal plasticity in the blind. Experimental techniques emphasized are in-situ electrophysiology and functional magnetic resonance imaging.

Date: May 18, 2011
Speaker: Olli-Pekka Kahilakoski (CoE, BECS)
Title: Bayesian analysis of sickness absence

Abstract: Individual factors associated with sickness absence have previously been studied using generalized linear models. With Bayesian analysis, we compared generalized linear models to Gaussian process models, which are flexible non-linear regression models that allow local changes in the response surface structure. Gaussian process models were found superior for predicting sickness absence with health questionnaire data in a sample of employees of a Finnish company.

We also did variable selection for Gaussian process models using Bayesian multiple comparisons. In agreement with previous studies, we found that depression and pain-related impairment at work are associated with increased sickness absence, with a possible saturation effect for depression.

Date: May 11, 2011
Speaker: Riku Linna, Ph.D. (CoE, BECS)
Title: Computational research on biopolymer systems

Abstract: In the research of biological systems, computational methods that are able to encompass both the detailed dynamics of the objects under investigation and the dynamics of the solvent in which they are immersed are of utmost importance. Through making measurements on and statistical analysis of the biological systems simulated using such methods and reflecting the observations on experimental findings new understanding is gained on nano and micron scale systems on which direct experimental observations are unattainable. I will introduce some of our research on biopolymers and the involved methods and outline the ongoing and future work.

Date: May 4, 2011
Speaker: Pekka Marttinen, Ph.D. (CoE, BECS)
Title: Detecting recombination events in bacterial genomes

Abstract: Recombination is recognized in bacteria as an important evolutionary force by which segments of DNA in the recipient organism are replaced by foreign segments from a donor cell. Consequently, the genome of a bacterium may consist of consecutive blocks obtained from different ancestors. Apart from complicating the definition of the species concept in bacteria, recombination may change the phenotype of a bacterium by giving birth to bacteria with increased virulence or antibiotic resistance. As a result, the effect of recombination on the evolution of bacteria as well as on the health of humans may be considerable. The purpose of our research is to develop novel Bayesian statistical methodology to allow more accurate and practically feasible identification of recombinogenic fragments within complete bacterial genomes from data sets comprising hundreds or thousands of bacterial samples. We accomplish this goal by describing the data in terms of a changepoint clustering model, where the detected clusters represent distinct evolutionary lineages and the changepoints represent locations in the genome where recombination events have locally modified the underlying clustering.

Date: April 27, 2011
Speaker: Jaakko Nieminen, M.Sc. (MEGMRI, BECS)
Title: MEGMRI: Novel brain imaging

Abstract: In MEGMRI project, we are developing a novel brain imaging device. For three years, we have aimed at a hybrid device capable for both magnetoencephalography (MEG) and magnetic resonance imaging (MRI). With MEG, one can measure the brain activity with millisecond time resolution; MRI offers means to image the structure of the brain with millimeter accuracy. A hybrid MEG-MRI device could offer improved source localization for MEG and also other benefits over separate MEG and MRI systems. The sensors measuring the magnetic fields of the neuronal activity for MEG collect also the MRI data. For this to be possible, MRI has to be performed in fields that are orders of magnitude lower than those of the conventional MRI; in our system, we perform MRI in a field comparable to the Earth?s magnetic field. We have modified a commercial multichannel MEG system such that it can be used also for MRI. In this talk, I will tell more about the MEGMRI project and the MEG-MRI system, show our first results, and discuss our future plans.

Date: April 20, 2011
Speaker: Gerardo Iniguez, M.Sc. (CoE, BECS)
Title: A brief introduction to complex networks

Abstract: From neural networks and metabolic interactions to the World Wide Web, scientific collaborations and the human society as a whole, complex networks can be found both inside us and around us. This makes their study an inherently interdisciplinary field that has drawn ideas from mathematics, physics, social sciences and biology, to name a few. The basic assumption behind the complex networks approach is that many phenomena in Nature can be feasibly represented as interactions (links) between a set of basic elements (nodes), where the interactions sometimes vary in time. Due to this universal presentation the metrics and computer algorithms used to study the dynamics of human relations, for example, can be applied to analyzing the network of protein-protein interactions. In this talk I intend to give a brief yet informative overall view of the successes and challenges of the field. I will start by citing a few common examples of empirical networks, followed by the basic mathematics of network theory, and end by introducing some useful tools for modeling both the network structure and the processes that take place on them, as well as the interplay between the two.

Date: April 13, 2011
Speaker: Roman Volinsky, Ph.D. (HBBG, BECS)
Title: Oxidized phosphatidylcholines reconstitute phospholipid "scramblase" activity in liposomes

Abstract: Lipid asymmetry is a ubiquitous property of the lipid bilayers in cellular membranes and its maintenance and loss play important roles in cell physiology, such as blood coagulation and apoptosis. The resulting exposure of phosphatidylserine (PS) on the outer surface of the plasma membrane has been suggested to be caused by a specific membrane enzyme, "scramblase". In spite of extensive research the role of "scramblase" in apoptosis has remained elusive. Here we show that lipid scramblase activity can be reconstituted in liposomes by oxidatively modified phosphatidylcholines. Combination of fluorescence spectroscopy and molecular dynamics simulations reveals that the mechanistic basis for this property of oxidized phosphatidylcholines is due to major changes imposed by the oxidized phospholipids on the biophysical properties of lipid bilayers, resulting in a fast cross bilayer diffusion ("flip-flop") of membrane phospholipids and loss of lipid asymmetry, requiring no "scramblase" protein.