Progress on monitoring and measuring brain activities

[johnnynanjiang]

1. Paradromics

Just tumbled upon Paradromics https://www.paradromics.com recently, and would like to share it with the community to get your thoughts around would it address some limitations mentioned in the book of “A Thousand Brains” about monitoring and measuring brain activities.

2. Also would like to see if other progresses or products are being made in this field.

3. FYI, below is a matrix generated by ChatGPT

Where Modern Neurotech Fits (2026 view)

Method / Product	Spatial Precision	Temporal Precision	Scale (# neurons)	Invasiveness	Key Idea
fMRI	Medium	Slow (~1s)	Whole brain	Non-invasive	Blood flow proxy
EEG (e.g. Emotiv)	Low	High	Whole brain	Non-invasive	Surface electrical
Kernel (Flow)	Medium	Medium	Medium	Non-invasive	Optical brain imaging
Neuralink	High	High	Hundreds–thousands	Invasive	Flexible high-density electrodes
Paradromics	High	High	~1,000+ channels	Invasive	High-bandwidth neural decoding
Blackrock Neurotech	High	High	Hundreds	Invasive	Utah array (clinical standard)
Synchron	Medium	High	Low–Medium	Minimally invasive	Stent-based (via blood vessels)
Precision Neuroscience	Medium–High	High	Medium	Surface implant	Thin cortical sheet
Single-neuron lab recording	Very high	Very high	Tiny	Invasive	Gold standard

[NeuroJoe]

Functional biocompatibility is one of the big problems that has yet to be solved by any BCI:

#neuroscience #neurotechnology #braincomputerinterface #neuralengineering #bioengineering #functionalperformance | TK Kozai .

One of Dr. Kozai’s review papers: Recent advances in neural interfaces—Materials chemistry to clinical translation | MRS Bulletin | Cambridge Core

[nleadholm]

Thanks @johnnynanjiang , looks interesting. It’s hard to get a sense of how high-fidelity the recording is. The ChatGPT table you included mentions “~1,000+ channels”, but this page suggests there are 421 micro-electrodes. Given the size of the device (it looks fairly big), it’s hard to know how precise the read-out from individual neurons will be. My guess is that it is sufficient for their purposes - i.e. a commercial product for people who otherwise can’t verbally communicate - but it’s hard to know whether it could be used to record from the 10-100k neurons in a column to observe distributed encoding as predicted by the TBT.

It’s still interesting to hear about developments in this space though, so thank you for sharing. If you’re curious about this topic, you might want to check out the work from Alipasha Vaziri. He’s been developing some really interesting optical techniques to address some of the limitations in the field.