Storage for free: a surprising property of a simple gain-control model of motion aftereffects

Abstract

If a motion aftereffect (MAE) for given adaptation conditions has a duration T s, and the eyes are closed after adaptation during a waiting period t w ¼ T s before testing, an unexpected MAE of a 'residual' duration T rT s is experienced. This effect is called 'storage' and it is often quantified by a storage factor r ¼ T rT =T , which can reach values up to about 0.7-0.8. The phenomenon and its name have invited explanations in terms of inhibition of recovery during darkness. We present a model based on the opposite idea, that an effective test stimulus quickens recovery relative to darkness or other ineffective test stimuli. The model is worked out in mathematical detail and proves to explain 'storage' data from the literature, on the static MAE (sMAE: an MAE experienced for static test stimuli). We also present results of a psychophysical experiment with moving random pixel arrays, quantifying storage phenomena both for the sMAE and the dynamic MAE (dMAE: an MAE experienced for a random dynamic noise test stimulus). Storage factors for the dMAE are lower than for the sMAE. Our model also gives an excellent description of these new data on storage of the dMAE. The term 'storage' might therefore be a misnomer. If an effective test stimulus influences all direction tuned motion sensors indiscriminately and thus speeds up equalization of gains, one gets the storage phenomenon for free.

The surface and deep structure of the waterfall illusion

“…The phenomenon has long been known, and it continues to present novel properties (van de Grind, van der Smagt, & Verstraten, 2004). One of the novel features of MAEs is that they can provide an ideal visual assay for distinguishing local from global eVects.…”

“…One of the features of MAEs that has posed problems for bottom-up accounts has been the phenomenon of storage (van de Grind et al, 2004). Wohlgemuth (1911) showed that the MAE will last for longer if there is a dark period immediately after adaptation.…”

Visual motion detection sensitivity is enhanced by an orthogonal motion aftereffect

“…Grunewald and Lankheet (1996) introduced broadly tuned inhibition among motion directions in the distributionshift model and attributed the unidirectional MAE seen in the orthogonal direction to inhibition resulting from the adapting stimuli that moved in opposite directions. Recently, van de Grind and colleagues developed a more elaborate computational model of motion adaptation in which divisive feed-forward inhibition was introduced as a gain-control mechanism (van de Grind, Lankheet, & Tao, 2003;van de Grind, van der Smagt, & Verstraten, 2004).…”

Visual Pattern Adaptation in Subjects with Photoparoxysmal EEG Response: Evidence for Increased Visual Cortical Excitability

“…Therefore, the MAE needed to be stored a comparatively longer time in the PPR group. To evaluate the resilience of the network changes that build up during adaptation, a range of storage times could be used to determine, for each group, whether the MAE duration plateaus or continues to decline with storage times that extend beyond their immediate MAE duration (after van de Grind et al 36 ).…”