μblog: engineering from the trenches

Here is the M1817A11 GSM/EDGE transceiver (I think!) from Infineon. I am way late for a trip so more text might come soon. Comments all-ways welcome.

infm1817a11-4x-stitched.JPG

del.icio.us | digg

I spend the first half of today participating in an EEG Mu/Beta rhythm experiment which tries to differentiate intent for left and right hand movement based on a four electrode montage. This was the first time that I did such a thing and also the first time that my own EEG was recorded. Although there is plenty of optimism around using EEG recording to control video games, playing Pong (video) is the state of the art. Companies like g.Tec already sell hardware ensembles that have enough processing power for at least rough, single-axis position interpretation. The trouble is that there is a lot of training involved and there is still the propensity for misclassification if the data is analyzed in real-time movement tasks where there is intent but no well-defined success criteria. I am hoping to do some more BCI experiments and will hopefully have some more things to report. Below is a plot of the results of my best trial where I had to think about moving either my left or right arm as per computer instructions. Please note that the classification error is down to almost 8% between 4.5 and 5 seconds after experiment initiation with the instruction being given between 2 and 4 seconds. Since this is technically a single-axis discrimination, I think I may be ready to play some neuro-pong!

del.icio.us | digg