[gus_massa]

Copy of a comment I made in a previous submission, a few hour ago: https://news.ycombinator.com/item?id=8207152

Well, the data is very noisy. The main problem is that this data doesn't have a before/after comparison. Is the 850nm light visible now or it was always visible???

It's also very difficult to make a fair comparison. The room must be the same, the light sources must be the same (a new coffeepot with a small led can ruin the experiment, removing a coffeepot because it has recently broken can ruin the experiment).

For a preliminary experiment, the before-after comparison is enough. For a serious experiment you need many voluntaries, compare the before-after signals of them all at the same time in the same experimental conditions, and double blind testing.

There is a small possibility that they are measuring "excitement" instead of light. The subject hears that they are now going to test with very near infrared light. He got exited. They measure that. Perhaps the flash makes a slight sound, perhaps the light operator makes a slight sound. (Perhaps the 850nm flash makes a sound that the other flashes don't make?)

[lotsofmangos]

If you read the text, it tells you that the 850nm was always visible.

[gus_massa]

I reed the previous version of this URL. IIRC it has the same graphs, but the text was much shorter, and didn't explain that. Replace 850 with 950 in my comment. The article now says that 950nm was not visible, but they don't link to the graph.