it actually isn't math that's needed, it's putting a load on the working memory.
so trying to keep a list of words in your head, or a series of numbers, or do math, or make music, etc., any of those would work.
this has been done, with a tongue-electrode delivering sonar sense to the brain - it's smart enough to map the signals into spatial knowledge pretty quickly.
i first read about it as a military technology, but here is a page on variations on the same theme:
http://www.artificialvision.com/sensub.htm
excellent point.
fyi, neurofeedback devices are pretty slick these days. i've used many that provide rich feedback from EEG or blood perfusion dynamics (passive or near infra red).
neural input definitely occurs via this feedback- you can make dramatic electrophysiological changes in the brain (like eliminating patterns that suggest ADHD) in about 10 hours of total training time.
that's not direct neural input, of course - the rich feedback i'm referring too is light/sounds, animations, dvd movie playback (with variable brightness / sound), and lots of games where control over the game is changed based on threshold EEG / HEG values.
there is one system that DOES provide a direct electromagnetic feedback on the EEG- it's called LENS: Low Energy Neurofeedback System. it uses micro-potentials to entrain the underlying EEG briefly, while sampling on many channels to provide baselines for feedback.
with an array of sensors the spatial resolution won't be that much worse than EEG, as it can employ the same kind of inverse solution algorithms to decompose the 3D infra red signal into regional or logical contribution to the blood flow changes. e.g. principal component analysis over partial differentials.
and i agree with you about your bandwidth observations. there may be ways to slave software/hardware into the EEG generation system, e.g. microcolums and macrocircuits, but we don't have a perfect idea of how those currently operate.
if we did it might be possible extract a denser information signal by integrating the variables of coherence, comodulation, phase, amplitude, frequency, etc., there is a LOT of information at any one spot on the scalp:) translating this back into observable behavior (for example, wheelchair movement) wouldn't be perfect at first but would encourage plasticity via feedback and which should remap cortex just the way we remap cortex daily, when learning new motor patterns (practicing piano, for instance).
well, the BOLD response measured via MRI peaks in several seconds, but you can get faster response with near infra red or passive infra red measurement. that's essentially what they are doing, i believe.
i use similar sensors for biofeedback - e.g. training perfusion dynamics in the frontal lobe.
it has great results, and you can watch the measurements fluctuate with concentration.
what hitachi has done new is the dense array and software thresholding the signals appropriately and controlling the train as "feedback". i'm sure they are using a moving window to sample the signal too; hemoencephalic dynamics are pretty variable.
ReGenesis, that great canadian cable series about a team of maverick scientists doing cutting edge microbio amid global intrigue:)
They had a story arc the 2nd half of last season, iirc - a cell line used typically had a hidden retrovirus that was millions of years old.
Worth checking out when it returns in the spring for it's 3rd season.
http://www.regenesistv.com/
Slashdot Mirror
it actually isn't math that's needed, it's putting a load on the working memory. so trying to keep a list of words in your head, or a series of numbers, or do math, or make music, etc., any of those would work.
this has been done, with a tongue-electrode delivering sonar sense to the brain - it's smart enough to map the signals into spatial knowledge pretty quickly. i first read about it as a military technology, but here is a page on variations on the same theme: http://www.artificialvision.com/sensub.htm
excellent point. fyi, neurofeedback devices are pretty slick these days. i've used many that provide rich feedback from EEG or blood perfusion dynamics (passive or near infra red). neural input definitely occurs via this feedback- you can make dramatic electrophysiological changes in the brain (like eliminating patterns that suggest ADHD) in about 10 hours of total training time. that's not direct neural input, of course - the rich feedback i'm referring too is light/sounds, animations, dvd movie playback (with variable brightness / sound), and lots of games where control over the game is changed based on threshold EEG / HEG values. there is one system that DOES provide a direct electromagnetic feedback on the EEG- it's called LENS: Low Energy Neurofeedback System. it uses micro-potentials to entrain the underlying EEG briefly, while sampling on many channels to provide baselines for feedback.
with an array of sensors the spatial resolution won't be that much worse than EEG, as it can employ the same kind of inverse solution algorithms to decompose the 3D infra red signal into regional or logical contribution to the blood flow changes. e.g. principal component analysis over partial differentials. and i agree with you about your bandwidth observations. there may be ways to slave software/hardware into the EEG generation system, e.g. microcolums and macrocircuits, but we don't have a perfect idea of how those currently operate. if we did it might be possible extract a denser information signal by integrating the variables of coherence, comodulation, phase, amplitude, frequency, etc., there is a LOT of information at any one spot on the scalp :) translating this back into observable behavior (for example, wheelchair movement) wouldn't be perfect at first but would encourage plasticity via feedback and which should remap cortex just the way we remap cortex daily, when learning new motor patterns (practicing piano, for instance).
well, the BOLD response measured via MRI peaks in several seconds, but you can get faster response with near infra red or passive infra red measurement. that's essentially what they are doing, i believe. i use similar sensors for biofeedback - e.g. training perfusion dynamics in the frontal lobe. it has great results, and you can watch the measurements fluctuate with concentration. what hitachi has done new is the dense array and software thresholding the signals appropriately and controlling the train as "feedback". i'm sure they are using a moving window to sample the signal too; hemoencephalic dynamics are pretty variable.
ReGenesis, that great canadian cable series about a team of maverick scientists doing cutting edge microbio amid global intrigue :)
They had a story arc the 2nd half of last season, iirc - a cell line used typically had a hidden retrovirus that was millions of years old.
Worth checking out when it returns in the spring for it's 3rd season.
http://www.regenesistv.com/