Technically yes. But a machine must be able to detect microsleep within a reasonable time frame for it to be useful as a detection device. But it is arbitrary as similar numbers for 5 second time frame would mean great degree of correlation with EEG. Similar percentages for 5 minute time frame would make this much less useful device, for example. So my argument still stands but it does depend on how they go about measuring this. for example, are each microsleep similar in character and intensity or very different. Also, how frequent are these microsleep and how much earlier in the sleepiness cycles do they emerge and how close are they to really falling asleep. This data could mean they were detecting microsleep a full hour before test subjects actually fall asleep which would make it literally INCREDIBLE!!!!!!!
Market Data 
Markets 
AI
