7 ms·

Human Brain Processing Speed clocked at less than 60 bits/s using Hicks Law

23 points by jonmc12 17y ago

bdfh42 17y ago
I don't know about you but my visual systems are processing at a greater bandwidth than that. Plus I am pretty sure that my hearing system is coping well with the simultaneous audio output of an MP3 track. Plus I am supposedly doing some work - I think that 60 bits/sec is rather under estimating the total.
- javanix 17y ago
  I think the 60 bits/sec figure is more analogous to encoding an mp3 than listening to one. Making distinct choices like you're probably doing with your work is the sort of processing that the article was referring to, not the "parsing" done by reading or listening to an mp3.
  - bdfh42 17y ago
    That's just picking one type of process running on what seems to be a massively parallel machine and saying that's the rate of the processor.
- sp332 17y ago
  FTA: "Martin uses his method to determine how much information the brain can process during lexical decision tasks. The answer? No more than about 60 bits per second." So, it's not 60 bits/sec of information processing, but 60 yes/no decisions per second. That's pretty fast.
  - bdfh42 17y ago
    And as each decision requires multiple inputs that would again rubbish the headline figure - total nonsense.
javanix 17y ago
This is pretty cool - as good as an example of the power of massive parallelism as I've ever seen, anyway.
jonsen 17y ago
As far as I remember the book The User Illusion: Cutting Consciousness Down to Size makes the measure to 16 bits/sec. A very interesting read now long ago. I must read it again. http://www.amazon.com/User-Illusion-Cutting-Consciousness-Penguin/dp/0140230122/ http://www.amazon.com/User-Illusion-Cutting-Consciousness-Pe...
newsdog 17y ago
I thought this might be the case. I figure that working intelligence, artificial or not, must take in a whole pile of data and throw out almost all of it, making it's decision based on a few bits. Thing is, the throwing out part is very elaborate and takes a lot of work.
- pcarmichael 17y ago
  I just finished listening to the audio book format of On Intelligence by Jeff Hawkins: http://www.amazon.com/Intelligence-Jeff-Hawkins/dp/0805078533/ http://www.amazon.com/Intelligence-Jeff-Hawkins/dp/080507853... (Hopefully I get this right... was pretty sleep deprived while listening to it.) In the book he describes the core part of intelligence as prediction - our brain is constantly making predictions from sensory inputs. While our brains do take in huge amounts of data, Hawkins' theory suggests that it is being operated on in massive parallelism - with different predictions happening simultaneously. Thus it isn't so much that all that data is being thrown away, as it is all being checked against known patterns from previous memories. He lays out reasoning for how intelligence, creativity, etc. are formed on top of this prediction model. It is a very interesting read.
radu_floricica 17y ago
Speech evolved comparatively late. I'd be willing to bet a motorcycle driver processes data a lot faster.

23 points by jonmc12 17y ago

I don't know about you but my visual systems are processing at a greater bandwidth than that. Plus I am pretty sure that my hearing system is coping well with the simultaneous audio output of an MP3 track. Plus I am supposedly doing some work - I think that 60 bits/sec is rather under estimating the total.

javanix 17y ago
I think the 60 bits/sec figure is more analogous to encoding an mp3 than listening to one. Making distinct choices like you're probably doing with your work is the sort of processing that the article was referring to, not the "parsing" done by reading or listening to an mp3.
- bdfh42 17y ago
  That's just picking one type of process running on what seems to be a massively parallel machine and saying that's the rate of the processor.
sp332 17y ago
FTA: "Martin uses his method to determine how much information the brain can process during lexical decision tasks. The answer? No more than about 60 bits per second." So, it's not 60 bits/sec of information processing, but 60 yes/no decisions per second. That's pretty fast.
- bdfh42 17y ago
  And as each decision requires multiple inputs that would again rubbish the headline figure - total nonsense.

javanix 17y ago

This is pretty cool - as good as an example of the power of massive parallelism as I've ever seen, anyway.

jonsen 17y ago

As far as I remember the book The User Illusion: Cutting Consciousness Down to Size makes the measure to 16 bits/sec. A very interesting read now long ago. I must read it again. http://www.amazon.com/User-Illusion-Cutting-Consciousness-Penguin/dp/0140230122/ http://www.amazon.com/User-Illusion-Cutting-Consciousness-Pe...

newsdog 17y ago

I thought this might be the case. I figure that working intelligence, artificial or not, must take in a whole pile of data and throw out almost all of it, making it's decision based on a few bits. Thing is, the throwing out part is very elaborate and takes a lot of work.

pcarmichael 17y ago
I just finished listening to the audio book format of On Intelligence by Jeff Hawkins: http://www.amazon.com/Intelligence-Jeff-Hawkins/dp/0805078533/ http://www.amazon.com/Intelligence-Jeff-Hawkins/dp/080507853... (Hopefully I get this right... was pretty sleep deprived while listening to it.) In the book he describes the core part of intelligence as prediction - our brain is constantly making predictions from sensory inputs. While our brains do take in huge amounts of data, Hawkins' theory suggests that it is being operated on in massive parallelism - with different predictions happening simultaneously. Thus it isn't so much that all that data is being thrown away, as it is all being checked against known patterns from previous memories. He lays out reasoning for how intelligence, creativity, etc. are formed on top of this prediction model. It is a very interesting read.

radu_floricica 17y ago

Speech evolved comparatively late. I'd be willing to bet a motorcycle driver processes data a lot faster.