Table 2 Levels of neuro-coupling and... | Emerald Publishing

Table 2

Levels of neuro-coupling and their recursive-capture potential

Level	Architecture	Human position	Typical illustration	Recursive-capture potential
0: external interaction	Behavioural input and output cross a conventional interface; no neural signal enters the loop	External user	Recommender system or manually controlled robot	Not distinctive to neuro-coupling
1: neural read-out	Neural activity is decoded, but system output does not recurrently modify the neural state used for subsequent control	Neural signal source, largely external to adaptation	One-off diagnostic electroencephalography (EEG) or non-adaptive neural classification	Low; privacy and misclassification remain salient
2: recurrent neuro-coupling	Decoded activity produces feedback that can alter subsequent user signals, but the machine mapping is fixed or changes only offline	Participant within a recurrent loop	Cursor or prosthesis control with performance feedback	Moderate and contingent on feedback, duration and dependence
3: adaptive co-coupling	User and machine update in response to one another during continuing operation; the system models the operator to optimise future interaction	Co-adapting constituent of the loop	Adaptive closed-loop BCI, neuroadaptive interface or responsive stimulation	Highest structural potential, but not an inevitable harm

Level	Architecture	Human position	Typical illustration	Recursive-capture potential
0: external interaction	Behavioural input and output cross a conventional interface; no neural signal enters the loop	External user	Recommender system or manually controlled robot	Not distinctive to neuro-coupling
1: neural read-out	Neural activity is decoded, but system output does not recurrently modify the neural state used for subsequent control	Neural signal source, largely external to adaptation	One-off diagnostic electroencephalography (EEG) or non-adaptive neural classification	Low; privacy and misclassification remain salient
2: recurrent neuro-coupling	Decoded activity produces feedback that can alter subsequent user signals, but the machine mapping is fixed or changes only offline	Participant within a recurrent loop	Cursor or prosthesis control with performance feedback	Moderate and contingent on feedback, duration and dependence
3: adaptive co-coupling	User and machine update in response to one another during continuing operation; the system models the operator to optimise future interaction	Co-adapting constituent of the loop	Adaptive closed-loop BCI, neuroadaptive interface or responsive stimulation	Highest structural potential, but not an inevitable harm

Source(s): Author’s own work