Blindsight and residual vision: How reliable is the redundant target effect as a diagnostic tool?

This review evaluates the diagnostic quality of the redundant target paradigm (RTP) as a tool for assessing residual visual capacities, like blindsight, in patients with visual field defects following brain injury. The RTP is based on the redundant target effect (RTE), whereby reaction times are slower to a single than to two redundant targets. In patients, the RTE indicates residual vision when the redundant target is presented within the blind field. By synthesizing data from the reviewed studies, we estimated measures of diagnostic quality. The specificity estimate was 89% (95%CI [69%, 97%]), indicating false-positive results were infrequent. Thus, a significant RTE suggests residual vision with high probability. However, sensitivity ranged from 42% (95%CI [19%, 68%]) to 77% (95%CI [57%, 90%]), showing that the RTP frequently fails to detect residual vision. The estimated reliability was low, with only 36% (95%CI [15%, 65%]) of cases demonstrating stable positive results. Therefore, absence of a significant RTE does not reliably indicate absence of residual vision. These findings limit the RTP’s suitability to draw conclusions about specific neuronal structures or functions. Moreover, several patients exhibited blindsight in other paradigms but not in the RTP. Consequently, RTP-results cannot serve as a prerequisite for interpreting other blindsight tasks. The RTP’s advantages, e.g., its ability to assess residual vision indirectly through a reaction-time effect, thereby avoiding biased response criteria, should encourage further research aimed at developing RTP-versions with improved diagnostic quality. However, low sensitivity and low reliability of current RTP-versions substantially limit its applicability for blindsight testing. • Reaction times are shorter for two redundant targets compared to a single target • This redundant target effect has been used to diagnose blindsight • Our review shows high specificity, i.e., a significant effect indicates blindsight • In contrast, low sensitivity and low reliability limit its applicability • Absence of a significant effect does not reliably indicate absence of blindsight