We have presented a Hebbian learning rule for spike-processing neural networks. The learning rule enables a network to learn associatively features of an object and to adapt synchronizing connections for object segregation. The learning process is performed with a self-organized switching between the learning and the recall mode. Known patterns are immediately recognized and new patterns are learned. We showed the properties of the learning rule by simulation examples for spatial and spatio-temporal patterns. For the processing of spatio-temporal patterns we chose a short term memory which is constructed of two cascaded leaky integrators. The point of the presented simulation examples was to show that the learning rule fulfills essential requirements for the application of spike-processing to machine vision. Features of an object are learned and recognized by a network without being disturbed by high background activity. When features of two or more objects are presented simultaneously the network separates the various objects in the temporal domain and performs thereby a scene segmentation. For the processing of real world data we have to proceed to multi-layer networks with advanced feature detectors as input units. Finally, a major advantage of the learning rule is that it does not lead to high computation costs and can easily be integrated in our neurocomputer for Spike-processing networks (NESPINN), which is under development.Feature linking via stimulus-evoked oscillations: Experimental results from cat visual cortex and functional implication from a network model",A biologically motivated and analytically soluble model of collective oscillations in the cortex", Biol.Modelling perceptual grouping and figure-ground segregation by means of active reentrant connections", Proc. Natl.

Within the EurIPACS HIPIN topic a generic HIS/RIS-PACS interface will be designed, implemented and evaluated. It is generally agreed that integration with the HIS/RIS is essential for the acceptance of PACS in a clinical environment. An interface between HIS/RIS and PACS allows more efficient usage of both systems, better integration of data, better consistency checking on shared data and better security and error handling. Also the PACS performance is improved by using HIS/RIS information to steer the image migration within the PACS. In this paper the functional specifications of the interface are described. These specifications are based on descriptions of clinical radiodiagnostic procedures. The generic interface consists of a common part, and of site specific adapters. The common part is identical for all incarnations and performs message scheduling, processing and logging. The adapters are specific for each communication standard, e.g. ACR-NEMA or HL7, and for each hospital. The interface will be implemented at the radiology department of the Philipps University Hospital in Marburg (Germany) and at the orthopaedic and neuroradiology departments of the hospital of the Free University in Brussels (Belgium).

This paper describes the functional specifications of the HIS/RIS-PACS interface that is being developed within the scope of the HIPIN topic as part of the European AIM/EurIPACS project. This functional specifications characterizes among other things the generic requirements for the interface, the radiological procedures, the HIPIN data dictionary, and the generic messages across the interface. In this paper, which is based on the first HIPIN deliverable, some features of the technical design of the HIPIN interface also are discussed.

The objective of the EurlPACS/HIPIN topic is to realize a generic HIS/RIS-PACS interface. The advantage of the generic interface concept is that it minimizes the efforts to realize a new instantiation of the HIPIN interface for a specific hospital. For a new instantiation only minimal adaptations in the software of the HIPIN interface itself and of the local HIS/RIS and PACS systems are required.

A visual object ‘pops out’ against the background, if its local features have a high degree of perceptual coherence: the object’s local features are linked by the visual system into a perceptual whole. A neuronal mechanism for such preattentive ‘object definition’ is likely to be based on stimulus-induced synchronized activities, as observed in the visual system of the cat [l, 2]. Based on neurophysiological evidence we developed a model neuron with the following properties [7]: 1. Convergent feeding connections can directly activate the model neurons spike discharge. 2. Modulatory action of linking inputs onto feeding inputs provides mutual enhancement and synchronization of stimulus-specific responses. The linking effect is represented by the synchronization of the spike trains of model neurons coupled by linking connections. Spatial and temporal continuity of a stimulus, thus, can produce synchronization in those neural assemblies that are activated by such stimulus. The response of a model of two identical layers is similar to that of cortical assemblies in our neurophysiological observations: Stimulated input regions that share some stimulus features induce synchronized ensemble activities via the laterally projecting recurrent linking connections, even if the stimuli are moving and are composed of spatially separated subregions. In a second type of simulation the associative property of the linking network is supported by a further layer with a classical type of associative memory [10–12]. Presenting a stimulus pattern (which is similar to stored patterns), the degree of association to one of these patterns depends on the spatial overlap of the patterns and the strength of the local linking connections in the lower layer.

Our models of visual information processing are based on the hypothesis that synchronized activities of sensory neurons serve to define perceptual relations: the features represented by the synchronized neurons are assumed to be linked and, thus, integrated into a perceptual entity. Recently, we found stimulus-related synchronizations in cat visual cortex that could play such role. These results are presented in chapter 2, together with discussions of the following questions: 1. What are the visual situations where stimulus-related activities in the visual cortex do become synchronized? 2. Where and by which neural mechanisms are synchronizations generated? 3. What possible roles do the synchronizations play in visual processing?

We recently discovered stimulus-specific interactions between cell assemblies in cat primary visual cortex that could constitute a global linking principle for feature associations in sensory and m...

We recently discovered stimulus-specific interactions between cell assemblies in cat primary visual cortex that could constitute a global linking principle for feature associations in sensory and motor systems: stimulus-induced oscillatory activities (35-80 Hz) in remote cell assemblies of the same and of different visual cortex areas mutually synchronize, if common stimulus features drive the assemblies simultaneously. Based on our neurophysiological findings we simulated feature linking via synchronizations in networks of model neurons. The networks consisted of two one-dimensional layers of neurons, coupled in a forward direction via feeding connections and in lateral and backward directions via modulatory linking connections. The models' performance is demonstrated in examples of region linking with spatiotemporally varying inputs, where the rhythmic activities in response to an input, that initially are uncorrelated, become phase locked. We propose that synchronization is a general principle for the coding of associations in and among sensory systems and that at least two distinct types of synchronization do exist: stimulus-forced (event-locked) synchronizations support “crude instantaneous” associations and stimulus-induced (oscillatory) synchronizations support more complex iterative association processes. In order to bring neural linking mechanisms into correspondence with perceptual feature linking, we introduce the concept of the linking field (association field) of a local assembly of visual neurons. The linking field extends the concept of the invariant receptive field (RF) of single neurons to the flexible association of RFs in neural assemblies.

Based on our neurophysiological findings, we developed a neural network model that performs feature linking via modulatory interactions. In the linked state, the activities of the model neurons are highly synchronized. The model’s performance is demonstrated in an example of contrast-dependent region linking and in examples of region definition in the Craik-O’Brien-Cornsweet illusion.

Until recently, no known physiological mechanism could explain the transient linking of local visual features into global coherent percepts. The authors have found such principles at a 'computational stage' of preattentive processing in the cat's primary visual cortex. The results show that stimulus-evoked oscillations of local processing units, representing local visual features, are transiently locked into a common resonance state by an appropriate global stimulus. The experimental results suggest further that the activities of those neural groups that represent features that are to be linked in the current visual situation become synchronized. Synchronization was found among assemblies in a single cortex column (where similar local features are processed) and, more important, among remote assemblies of the same cortex area and even among assemblies in different areas when the assemblies share some common local visual features. These cortical linking features are considered to be the neural representations of the corresponding perceptual linking features.<>