Figure 2.
Closed-loop methods for measuring iso-response stimuli. (A) Closing the loop by tuning stimulus parameters according to measured responses. In response to a stimulus with two components s1 and s2 (top), a recorded neuron (right) responds with spikes that are automatically detected, for example by a threshold criterion (bottom). The spike response is then compared to a chosen target criterion (left), which may be the number of spikes, the timing of the first spike, the probability of spiking, or any other accessible response feature. According to this comparison, the values of s1 and s2 are adjusted for the next stimulus presentation in order to approach the target response. (B) Potential search strategy in radial directions of the stimulus space. This combines several linear searches, which can be performed sequentially or interleaved, typically starting near the origin so that overly strong stimulation is avoided. (C) Potential search strategy by tracking iso-response curves. Here, previously measured iso-response stimuli are used as starting conditions for searching nearby stimuli that yield the same response. This can be done, for example, by changing the ratio of s1 and s2 while keeping the same radial distance as for a previously measured iso-response stimulus and then tuning this radial distance until the desired response is obtained. As compared to the strategy of pure radial searches in (B), this sequential search can provide higher recording efficiency, but does not allow interleaving multiple searches.