Since January 2007 I am a PhD student at the university of Genova (Italy). I work under the supervision of prof. Sergio Martinoia in the Neuroengineering and Bio-Nanotechnology group. Currently I work at the California Institute of Technology with dr. Daniel Wagenaar.

Background. The title of my research proposal was "stimulation-induced plasticity of neuronal networks". We use neuronal cultures (in-vitro) as a generic model for neuronal systems in living organisms. The brain tissue is taken from rat embryo's so that it it still in a state where it is growing. We then remove all the structure that is present in the original brain in order to avoid effects due to the particular 'design' of a brain area. By themselves, the neurons then reconnect and form a new network. After roughly a week this network starts to exhibit very stereotypical activity.

The brain cells are now put on a glass dish with integrated electrodes. These electrodes can pick up the electrical signals the neurons use to communicate. Meanwhile I can also send electrical signals back through the same electrodes. How the networks respond to that stimulation is the central topic of my project.

Experiments and analysis. The experiments are in reality effect quite simple. Most time is spent on the analysis of the data. One particular problem with the electrical stimulation is that the response can be quite long, and as a result, responses to subsequent stimuli will interact. I devised an analysis technique to disentangle these two (or more) stimuli. With the experiments we try to better understand the interaction of two inputs that arrive almost simultaneously. This is a very low-level version of for instance organisms connecting the auditory and visual part of a single stimulus that arrive in close succession. It is also related to attentional effects, where a subject processes a stimulus more effectively if she is 'primed'.

Relevance. The experiments are currently directed at understanding how the networks respond to electrical stimulation as an experimental model for information processing in the brain; my electrical stimuli are a substitute for information that comes in from the senses. In this case I try to relate effects that I see from my stimulation to observation in behavioral experiments.

Understanding the effect of external stimulation also has a more applied goal. There are many efforts underway to uncover how we can interfere with the brain's information processing through external stimulation.

Publications*. The first results on the new analysis technique and response interactions are published in Physical Review E.

P.L. Baljon, Chiappalone, M. and Martinoia S. (2009) "Interaction of electrically evoked responses in networks of dissociated cortical neurons" in Physical Review E 80(3). [download PDF][pubmed]

Earlier results about effects of stimulation on ongoing activity were published for my oral presentation at the IEEE conference for Engineering in Medical and Biological Science (EMBS).

P. Massobrio, Baljon, P.L., Maccione, A., Chiappalone, M. and Martinoia, S. (2007) "Activity modulation elicited by electrical stimulation in networks of dissociated cortical neurons" in Proceedings of the 29th Annual International Conference of the IEEE EMBS [download PDF][pubmed]

Until december 2006 I worked as 'researcher' at the Artificial Intelligence department of the University of Groningen under prof. L. Schomaker. My primary (and professionally only) interest is Brain-Computer Interfacing. I use the EEG signals obtained from the scalp, to 'read the mind' of the participant. Having read the mind of the user, I can steer a cursor on the computer screen according to the intention of the user. As the user has control over her intentions, she has control over the cursor.

I work in the 'Moving Thoughts' group for BCI research in Groningen. This is a collaboration between Experimental Psychology (prof. R. de Jong), Human Movement Sciences (prof. B. Otten) and Artificial Intelligence (prof. L. Schomaker). More information about the group can be found at the site of Moving Thoughts.

There are numerous problems with this type of control; most importantly the amount of noise in the EEG signal. Therefore signal processing is a key element in this field of research. Furthermore I work with machine learning techniques from my background in Artificial Intelligence. Most importantly, I used Hidden-Markov Models which are able to encode time information in a signal. In my current work, I implemented more general pattern recognition tools to classify EEG data,

The hypothesis in my master thesis (PDF version, 1.6Mb) was that some of the variability in the EEG is actually not noise, but meaningful time structure in the signal. In speech for example, the time structure in the sounds we perceive is vital to distinguish the words 'law' and 'all'. This turned out to be less true for the time structure in EEG during BCI control. I presented these results in a poster (PDF version, 268Kb) at the Inaugural conference of the Society of Applied Neuroscience.

An introduction to our research is the 'lunch talk' I gave at the Artificial Intelligence group in Groningen on 14/12/06 about Brain-computer Interfaces (PPT version, 5.9Mb). It was an introductory talk, going into detail about the specific type of BCI research in Groningen (EEG-based BCI) and our specific setup (BCI2000).

Furthermore, the in my opinion key articles on this topic are the following:

• Curran & Stokes (2003), Learning to control brain activity: a review of the production and control of EEG components for driving brain-computer interface (BCI) systems, in Brain & Cognition 51(3):326-36
• Wolpaw & McFarland (2004), Control of a two-dimensional movement signal by a noninvasive brain-computer interface in humans, in Proceedings of the National Acedamy of Sciences 101(51):17849-54

A good introductory read (though not specifically for non-invasive) is the Nature article by the pioneer in invasive BCIs:

• Nicolelis (2001) Actions from Thoughts in Nature 409(6818):403-7

Of particular interest for my master thesis is the paper Obermaier et al. (2001) Hidden Markov models for online classification of single trial EEG data in Pattern Recognition Letters 22:1299-309.