TO TOP

FOR 2812-Kolloquium: The probabilistic nature of coding and using structured episodic memories

25.10.2024

Accumulating behavioral and neural evidence suggests that incoming sensory input is represented and combined with generalized semantic knowledge in a fundamentally probabilistic way during perceptual processes in the brain. Meanwhile, episodic memories are traditionally treated outside of this perceptual context despite the fact that semantic memories must emerge from episodes experienced earlier by the observer and that episodic memory is a significant part of the internal representation. Hence, it is not even known if episodic memories are also treated probabilistically, that is, whether they are encoded together with their uncertainty just as immediate sensory information is. To fill this gap, I will present a series of behavioural studies exploring the extent to which episodic memory pocesses characteristics of probabilistic coding. First, I will demonstrate that encoding of episodic memories are well-calibrated that is, they are recalled later from long-term memory together with their subjective uncertainty, indicating a probabilistic representation of memory details. Second, I will show how the encoding of a feature of episodic memory is systematically biased by the summary statistics of the feature that are also automatically extracted and stored during exposure. Third, I will present evidence about how instantaneous image configurations and semantic background in trials influences memory encoding. The results suggest that encoding and recall of long-term episodic memories follow the same probabilistic principles as perception. As a concequence, momentary perceptual input, semantic knowledge and individual episodic memory traces can be treated and combined in a fundamentally uniform manner in the brain.

Speaker:

  • Jozsef Fiser (Central European University)

Venue:
NB 3/57, Ruhr University Bochum

Starting time:
2:00 PM