Measurement back-action · Grey Matter

Measurement back-action is the change that the act of measuring makes in the brain being measured, and it limits every method that gets close enough to read single cells: the closer and longer the reading, the more the tissue reacts to the reader.

Measurement back-action. Measurement back-action is the change that the act of measuring makes in the brain being measured, and it limits every method that gets close enough to read single cells: the closer and longer the reading, the more the tissue reacts to the reader.

At the scale of neurons there is no passive observer. An electrode pushed into the cortex cuts vessels and cell processes on the way in, and the tissue then treats it as a foreign body: microglia and astrocytes gather around it over weeks and wall it off in a sheath, a glial scar, while nearby neurons may die back. Light used to excite fluorescent indicators or to drive optogenetic channels is partly absorbed and heats the tissue, and enough light alone changes firing. Fluorescent calcium indicators bind calcium and so slightly buffer the very signal they report. Even stimulation meant only to probe (a test pulse, an imaging session) leaves traces in a network that learns from what happens to it.

It degrades recordings over time. As the scar thickens and moves neurons away, implanted electrodes record fewer and smaller spikes; long-term stability is one of the main engineering goals of flexible, thinner probes.

It is managed by budgets. Light power, stimulation charge and implant size are kept below levels known to change the tissue, and control experiments (light without the opsin, for example) check what the method itself does.

Far from the brain it vanishes. EEG, MEG and fMRI do not measurably disturb neural activity, which is part of their price: their distance is also what makes them blurred.

Getting closer buys signal and costs disturbance.

The methods that read single neurons are exactly the ones whose presence the tissue notices and reacts to.

Questions: Can the light used in optogenetics change the brain it is meant to control? Yes, through heat. Part of the light delivered through a fibre is absorbed by tissue, and modelling and measurements show that light powers within the range used in experiments can warm the tissue near the fibre tip by around a degree or more, which in some conditions is enough to change firing on its own, in neurons that carry no opsin at all. Careful experiments therefore keep light power and duty cycle low and run the same light in animals without the opsin, so that any effect of light alone can be subtracted from the effect of switching the cells. When does measuring a brain become changing it, and which proposed right covers that? Every method close enough to read single neurons also acts on them: electrodes provoke a tissue response, light heats, and recordings used for feedback train the brain that produces them, as users of a brain-computer interface learn to drive its decoder. The proposed right to mental integrity, in the framework of Ienca and Andorno, protects against harmful interference with a person's neural processes, and the right to psychological continuity protects their sense of identity from changes they did not agree to. Applied to implants, these rights ask that the effects of the device on the brain, intended or side effects, be known, disclosed and reversible where possible.