A sensory preprocessing module for an algorithmic model of recognition memory Computational Neuroscience


Recognition memory refers to the ability to differentiate novel information from the one encountered before. At the level of subjective experience, it is believed to have two components: familiarity and recollection, illustrated by the so-called butcher-on-the-bus phenomenon (Mandler 1980). This refers to the confusing situation where a person feels confident to have seen a face before (familiarity) without being able to retrieve contextual details, e.g when or where they met (recollection). According to dual-process model of recognition memory, recollection is supported by the hippocampus, while familiarity mainly relies on the perirhinal cortex (Yonelinas 1994, Eichenbaum 2007). By contrast, single-process models suggest that the representations in the hippocampus and perirhinal cortex differ in memory strength instead of underlying processing (Squire et al. 2007). We have implemented an alternative model of recognition memory showing that the properties of both the input and the processing modules are sufficient to account for recognition performance without the involvement of distinct memory retrieval mechanisms.
Currently, images of letters are used as stimulus material, which is too simple for human subjects. We would like to use naturalistic inputs so that the predictions of our model can be directly tested in experiments. Two possible models may be integrated to our work to process naturalistic images. The first is a hierarchical model object recognition proposed by Riesenhuber and Poggio (1999). This model has previously applied by one of the lab members (Dr. Amir Azizi) and could potentially be implemented as the input module in our model. The second approach is to apply deep neural networks (DNN) to extract image features. The goal of this project is to implement these modules, integrate them as input preprocessing modules, and explore their influence on performance. The project requires knowledge of Matlab and Python. Experience with deep neural networks will be an advantage.

Eichenbaum, H., Yonelinas, A. and Ranganath, C. (2007). The Medial Temporal Lobe and Recognition Memory. Annual Review of Neuroscience, 30(1), pp.123-152.

Mandler, G. (1980). Recognizing: The judgment of previous occurrence. Psychological Review, 87(3), pp.252-271.

Riesenhuber, M. and Poggio, T. (1999). Hierarchical models of object recognition in cortex. Nature Neuroscience, 2(11), pp.1019-1025.

Squire, L., Wixted, J. and Clark, R. (2007). Recognition memory and the medial temporal lobe: a new perspective. Nature Reviews Neuroscience, 8(11), pp.872-883.

Yonelinas, A. (1994). Receiver-operating characteristics in recognition memory: Evidence for a dual-process model. Journal of Experimental Psychology: Learning, Memory, and Cognition, 20(6), pp.1341-1354.


Olya Hakobyan and Prof. Dr. Sen Cheng

The Institut für Neuroinformatik (INI) is a central research unit of the Ruhr-Universität Bochum. We aim to understand the fundamental principles through which organisms generate behavior and cognition while linked to their environments through sensory systems and while acting in those environments through effector systems. Inspired by our insights into such natural cognitive systems, we seek new solutions to problems of information processing in artificial cognitive systems. We draw from a variety of disciplines that include experimental approaches from psychology and neurophysiology as well as theoretical approaches from physics, mathematics, electrical engineering and applied computer science.

Universitätsstr. 150, Building NB, Room 3/32
D-44801 Bochum, Germany

Tel: (+49) 234 32-28967
Fax: (+49) 234 32-14210