Rewrite Science - The Universal Semantic Calculus and Grammatical Cosmos
Symposium 10A - CASYS 07

The "Time-Loop" Model of Visual Perception
Uwe Kaempf
Department of Psychology, Dresden University of Technology, D-01062 Dresden
+49-(0)3"51-46333]79 (phone/fax) uwe.kaempf@psychologie.tu-dresden.de

Abstract

As an introspective observer from the cockpit of our ego, we are aware of an outlook into the visually perceived environment of a spatially extended holistic representation with simultaneous awareness of all its visible parts. Nonetheless, visual perception transforms spatial extension into temporally extended codes, being registered, e.g., as an oscillatory EEG pattern of successive phases or stages of brain activation. Since this kind of temporal extension of perceptual processing is in a rather fundamental conflict with the real-time comprehension of ongoing events, the hypothesis of a "time-loop" model of visual perception is presented here in order to handle the contradiction.

The functional architecture of recurrent visual processing is organized into a hierarchy of interdependent levels, the timing of which is known to be coordinated by neuron loops lasting up to more than 500 msec of temporal delay, the so-called synfire chains. In spite of this. these neuronal loops are synchronized with a precision of less than 1 msec in their repetitive re-occurrence of activating one and the same neuron. Obviously, if the actual visual input is fed into such a network, its processing results are retarded with respect to the changing environmental situation. Thus, the visual signal's recurrent processing is critically dependent on advanced predictions anticipating the next input in order to adjust its riming aspects in a way which stabilizes the equilibration of non-linear brain dynamics. In order to get a synchronous visual flow in both left and right brain hemispheres, as is needed to account for the coherence e.g. of binocular stereo vision, the temporally delayed information of the left and right eyes' visual fields, which are processed in different hemispheres, should be integrated with anticipated versions of their complements in order to close me time gap.

Therefore, the "time-loop" model is presented here as a quantum mechanical approach in order to account for some of the mentioned phenomena. From this point of view, the issues of visual processing under discussion may be reinterpreted in line with the hypothesis of an absorber theory for quantum brain dynamics. Based on the Wheeler-Feynman approach, the transactional interpretation of quantum mechanics had been originally proposed as an account for anomalous resonance coupling between spatio-temporally distant measurement partners in entangled quantum states of spatio-temporal non-locality, quantum teleportation etc, (Einstein-Podolsky-Rosen paradoxes).

Applied to quantum brain dynamics, this view seems applicable as an anticipative resonance coupling model for aspects of cortical synchronization and ideomotor action control. According to this model, recent findings about so-called mirror neurons in the brain cortex are suggested to be associated with temporal rather than spatial mirror functions of visual processing, due to phase conjugate adaptive resonance coupling known from non-linear optics. Therefore, it is proposed to consider the registered activation patterns of neuronal assemblies incorporated in synfire chains (in analogy to the mentioned EEC brain wave patterns of. e.g.. evoked potentials after visual stimulation etc.) not as a result of retarded brain communication processes. Rather these patterns are suggested to be better understood as a part of a standing waves system generated by a counterbalance between the actual input's delayed bottom-up processing and advanced anticipative signals from mirror neurons taking part m lop-down recurrent information processing.