Singing Gabors: Visual stimulus locking of EEG is modulated by temporal congruency of auditory stimuli

In this study, we used what we called back then the "Singing Gabors".
These consist of Gabor functions, where one of the parameters (here orientation parameter) is modulated either in a congruent or incongruent way with the audio signal.

Guess which of the following videos is the congruent one:

In the congruent case, the audio signal's frequency changes simultaneously with  the orientation of the Gabor. This leads (in my case) to a vivid illusion of agency, meaning that I perceive the animated Gabor function as if it was producing the sound itself (the paper deals with much simpler questions though).

In the incongruent case, the audio signal is orthogonal to the changes in the orientation parameter of the Gabor function. Still it is interesting to realize how human perception is biased for fusing these two as if they were congruent.