3 ms·
> As participants recalled the images, the receptive fields in the highest level of visual processing were the same size they had been during perception — but t
by smaddox 4y ago
> As participants recalled the images, the receptive fields in the highest level of visual processing were the same size they had been during perception — but the receptive fields stayed that size down through all the other levels painting the mental image. The remembered image was a large, blurry blob at every stage.
> [...]
> So depending on whether information is coming from the retina or from wherever memories are stored, the brain handles and processes it very differently. Some of the precision of the original perception gets lost on its way into memory, and “you can’t magically get it back,” Favila said.
This is probably more a comment on imprecise or confusing journalism than anything about the paper, but just because a larger number of early neurons were activated when recalling a memory doesn't mean the memory itself is blurry or low resolution. Perhaps the memory is just compressed (embedded) in such a way that decompressing it requires activation of a large number of early-stage neurons. This is basically how a UNet works, although UNet decoders don't (typically?) share neurons (weights) with the the encoder.
- nickelpro 4y agoYa that's the massive weakness of these studies. They always read more into their data than what's actually there. You're basically taking a snapshot of lights in a city at night. You can't see any of the people or what they're doing, but you see a lot of lights along the waterfront. You know a shipment of freight just came in, so you make the reasonable assumption they're unloading the freighter. But they could just as easily be dumping the tea in the harbor. There's no mechanical description of what is happening, just the observation that something appears to be happening.