5 ms·

This is interesting — could you link your research? And are these DC fields you're applying using electrodes across the cranium? Or AC using an external coil?

mattkrause 4y ago
Sure! If this is too long or scattered, I've got a review article for a "general audience" coming out soon in PLoS Biology. -- The family of techniques I work on is called "transcranial electrical stimulation", or tES. It involves placing electrodes (either saline-soaked sponges or silver disks coated with a conductive paste) on the user's intact scalp. Weak currents (< 4 mA) are passed between them to create electric fields inside the head. Various forms of tES are distinguished by the waveform of that current: using direct current is called transcranial direct current stimulation (tDCS); using alternating current is called tACS instead. This is really easy to set up: all you need is a current source and some conductors; you could probably get going for $15. However, it's been very unclear—and controversial—what it actually does. Some groups find giant effects on behavior and disease symptoms; others report that these don't replicate. There are lots of reasons this might happen, but one common suggestion is that the stimulation doesn't do anything at all! To sort this out, my colleagues and I have been recording neural activity from macaque monkeys receiving tES. Macaques are very similar to humans, and we use the same gear meant for people (mostly this one: https://www.neuroelectrics.com/solutions/starstim https://www.neuroelectrics.com/solutions/starstim). Under these conditions, we find that it does affect neural activity---but often not in the ways people expect. In our first study, we applied tDCS while the animals did a "foraging" task. We found that animals learned faster with stimulation. While it didn't cause massive changes in brain "excitability" (one hypothesis), stimulation changed how visual flowed from a sensory area to those involved in learning/memory. Here's the paper: https://pubmed.ncbi.nlm.nih.gov/29033331/ https://pubmed.ncbi.nlm.nih.gov/29033331/ This blog had a nice write-up: https://hackaday.com/2017/11/13/shockingly-darpas-brain-stimulator-might-not-be-complete-nonsense/ https://hackaday.com/2017/11/13/shockingly-darpas-brain-stim... That study was mostly aimed at seeing if tES did anything at all, but the single-neuron data was a bit muddled since the hypothesis wasn't too clear. We therefore switched to using alternating current, which was thought to "entrain" neurons so that they fire rhymically at the same frequency. We found that it did--and could even affect neurons in deep brain structures. Paper: https://pubmed.ncbi.nlm.nih.gov/30833389/ https://pubmed.ncbi.nlm.nih.gov/30833389/ One problem is that tACS drives nerve fibers in the skin, causing an itchy feeling near the electrodes. In theory, this could indirectly change neural activity as the resulting sensory activity propagates through the brain. To distinguish this from a "direct" effect of the electric field on the brain, we numbed the skin with topical anesthetic. We found that tACS had similar effects on neurons in both cases, suggesting that it does indeed directly affect the brain. Paper: https://pubmed.ncbi.nlm.nih.gov/33001971/ https://pubmed.ncbi.nlm.nih.gov/33001971/ So....if all that works, why is it so hard to use tACS to improve behavior? In our prior experiments, we had targeted conditions where ongoing brain activity was unstructured, the better to detect faint changes in spike timing. In human studies, however, tACS users mostly want to enhance an ongoing oscillation, where neurons are already firing in sync. Using a combination of experiments and modeling, we show that under these conditions, tACS and "natural" brain activity vie for control of spike timing, and the resulting competition can make neurons fire more or less rhythmically. I'm excited about this result because we do often want to break up brain oscillations (e.g., in epilepsy) and it explains why things have been so hard. Paper: https://pubmed.ncbi.nlm.nih.gov/35613140/ https://pubmed.ncbi.nlm.nih.gov/35613140/
- voldacar 4y ago
  Nice, I'll be paying attention. I've been planning on messing around with tdcs for a while now but I've never sat down and actually built the (admittedly really simple) device needed. Possibly a dumb question but do you have to use electrodes that require that gel substance for good conductivity, or are dry eeg electrodes sufficient? Or is this irrelevant if the signal comes from a current source?
  - mattkrause 4y ago
    Gel is probably the way to go. One usually specifies the stimulation in terms of current (1 mA, for example). You want to deliver that with as little voltage as possible and because V=IR, that means minimizing the electrode resistance. Minimizing voltage is good because the power delivered is proportional to V², and the power is what makes stimulation uncomfortable by heating/burning the skin (and I guess, the tissue below if you go really crazy). Most commercial systems have a voltage cut-off so they'll deliver (say) up to 2 mA but no more than 12 volts. The good news is that gel and saline are both very cheap. I bought $50 worth of SIGNAGEL in 2015 and the lab has hardly made a dent in it. Do shop around though, because some "wellness" places mark it way up.
- f3rnando 4y ago
  AIs might alredy be juggling with emw, we can still do the jerk, play the fool, use parables, breakbeat, GANs, etc. No AI gets a joke...

mattkrause 4y ago

Sure! If this is too long or scattered, I've got a review article for a "general audience" coming out soon in PLoS Biology. -- The family of techniques I work on is called "transcranial electrical stimulation", or tES. It involves placing electrodes (either saline-soaked sponges or silver disks coated with a conductive paste) on the user's intact scalp. Weak currents (< 4 mA) are passed between them to create electric fields inside the head. Various forms of tES are distinguished by the waveform of that current: using direct current is called transcranial direct current stimulation (tDCS); using alternating current is called tACS instead. This is really easy to set up: all you need is a current source and some conductors; you could probably get going for $15. However, it's been very unclear—and controversial—what it actually does. Some groups find giant effects on behavior and disease symptoms; others report that these don't replicate. There are lots of reasons this might happen, but one common suggestion is that the stimulation doesn't do anything at all! To sort this out, my colleagues and I have been recording neural activity from macaque monkeys receiving tES. Macaques are very similar to humans, and we use the same gear meant for people (mostly this one: https://www.neuroelectrics.com/solutions/starstim https://www.neuroelectrics.com/solutions/starstim). Under these conditions, we find that it does affect neural activity---but often not in the ways people expect. In our first study, we applied tDCS while the animals did a "foraging" task. We found that animals learned faster with stimulation. While it didn't cause massive changes in brain "excitability" (one hypothesis), stimulation changed how visual flowed from a sensory area to those involved in learning/memory. Here's the paper: https://pubmed.ncbi.nlm.nih.gov/29033331/ https://pubmed.ncbi.nlm.nih.gov/29033331/ This blog had a nice write-up: https://hackaday.com/2017/11/13/shockingly-darpas-brain-stimulator-might-not-be-complete-nonsense/ https://hackaday.com/2017/11/13/shockingly-darpas-brain-stim... That study was mostly aimed at seeing if tES did anything at all, but the single-neuron data was a bit muddled since the hypothesis wasn't too clear. We therefore switched to using alternating current, which was thought to "entrain" neurons so that they fire rhymically at the same frequency. We found that it did--and could even affect neurons in deep brain structures. Paper: https://pubmed.ncbi.nlm.nih.gov/30833389/ https://pubmed.ncbi.nlm.nih.gov/30833389/ One problem is that tACS drives nerve fibers in the skin, causing an itchy feeling near the electrodes. In theory, this could indirectly change neural activity as the resulting sensory activity propagates through the brain. To distinguish this from a "direct" effect of the electric field on the brain, we numbed the skin with topical anesthetic. We found that tACS had similar effects on neurons in both cases, suggesting that it does indeed directly affect the brain. Paper: https://pubmed.ncbi.nlm.nih.gov/33001971/ https://pubmed.ncbi.nlm.nih.gov/33001971/ So....if all that works, why is it so hard to use tACS to improve behavior? In our prior experiments, we had targeted conditions where ongoing brain activity was unstructured, the better to detect faint changes in spike timing. In human studies, however, tACS users mostly want to enhance an ongoing oscillation, where neurons are already firing in sync. Using a combination of experiments and modeling, we show that under these conditions, tACS and "natural" brain activity vie for control of spike timing, and the resulting competition can make neurons fire more or less rhythmically. I'm excited about this result because we do often want to break up brain oscillations (e.g., in epilepsy) and it explains why things have been so hard. Paper: https://pubmed.ncbi.nlm.nih.gov/35613140/ https://pubmed.ncbi.nlm.nih.gov/35613140/

voldacar 4y ago
Nice, I'll be paying attention. I've been planning on messing around with tdcs for a while now but I've never sat down and actually built the (admittedly really simple) device needed. Possibly a dumb question but do you have to use electrodes that require that gel substance for good conductivity, or are dry eeg electrodes sufficient? Or is this irrelevant if the signal comes from a current source?
- mattkrause 4y ago
  Gel is probably the way to go. One usually specifies the stimulation in terms of current (1 mA, for example). You want to deliver that with as little voltage as possible and because V=IR, that means minimizing the electrode resistance. Minimizing voltage is good because the power delivered is proportional to V², and the power is what makes stimulation uncomfortable by heating/burning the skin (and I guess, the tissue below if you go really crazy). Most commercial systems have a voltage cut-off so they'll deliver (say) up to 2 mA but no more than 12 volts. The good news is that gel and saline are both very cheap. I bought $50 worth of SIGNAGEL in 2015 and the lab has hardly made a dent in it. Do shop around though, because some "wellness" places mark it way up.
f3rnando 4y ago
AIs might alredy be juggling with emw, we can still do the jerk, play the fool, use parables, breakbeat, GANs, etc. No AI gets a joke...