It’s the first question I asked when I set out to design a tDCS device. I didn’t find any sources that were too accurate or particular.
There is some information out there on the galvanic resistance of the human body including on Wikipedia. Also it’s an interesting fact that galvanic skin resistance is highly dependent on sympathetic nervous system activation, can change rapidly, and is used as an important measurement in psychology and psychiatry experiments, in polygraph machines, and in Scientology auditing.
Bottom line is, the resistance of dry skin can be anywhere between 10 kOhm and 1 MOhm, while the resistance of the wet stuff inside us is more in the few hundred Ohms range. In AC stimulation like TENS, the impedance of dry skin is greatly reduced due to the AC nature of the applied electricity, because the electrode and the wet matter below the skin form a capacitor, which allows AC through while blocking DC. With tDCS, naturally we cannot rely on this.
To reduce the resistance of the dry outer skin layer, in medicine people usually use wet electrodes, like gel electrodes in EKG and EEG, and saline-soaked sponges for stimulation and iontophoresis. In fact, there is a good reason why most tDCS systems use wet sponges. It’s really simple, doesn’t cause a mess (not having to wash your hair is a plus), hair doesn’t really mess with the results, and it can handle Wattages that would heat up or break down gels.
While there’s a lot of talk about resistance of the electrodes, the fact is that if they aren’t dry, it’s safe to consider them a dead short. I’ve used sponge electrodes with an approximate area of 25 cm2 and with half as much at about 10 cm2, without any noticeable difference in the voltage dropped.
As for HD tDCS and gel electrodes, there was a research paper by HD tDCS pioneers Soterix Medical some time back, where they tested various electrode configurations. Most of them heated up, broke down under DC, their resistance skyrocketing, but their “winner” stayed below 10 Volts of drop throughout. In fact, I’m convinced a simple wet sponge fares just as well, or probably better, for a fraction of the price. In their case, there were other considerations in favor of gel electrodes &ndash like integration with existing QEEG positioning meshes and electrode bodies.
My experience is that for extracranial electrode configurations (ie. with one electrode placed on a shoulder, arm or otherwise not on the head), the total resistance of electrodes plus body tends to be 3500—4000 Ohms, while with both electrodes on the head, it’s more around 1000—2000 Ohms. These values are for quality sponge electrodes (Amrex) with a metal mesh backplate, and soaked (but not dripping) with standard .9% saline solution.
This means that for a tDCS device to provide enough juice for most scenarios, it has to be able to provide up to 8 Volts on its terminals. Most devices go somewhat above this, to deal with suboptimal connections, but my opinion is that for a DIY, or otherwise non-medical device, it’s better to stay around this value for sake of safety.