Event-driven spiking neutral network system for detection of physiological conditions

1 . An event-driven spiking neural network system implemented on a chip for detecting a physiological condition of a person based on a detected physiological signal of the person, comprising at least the following components:

a plurality of sensors, wherein each sensor is configured and arranged to detect a physiological signal and to convert the physiological signal in a sensor signal indicative of the physiological signal,

a plurality of signal conversion modules, wherein each is configured and arranged to receive one or more of the sensor signals from the plurality of sensors and to convert the sensor signals in at least one time series of discrete events by means of a Lebesgue transformation, wherein the generation of events is independent of a global synchronizing clock for synchronizing the components of the system, wherein foreach sensor signal received by the conversion module the signal conversion module generates at least two time series of events, wherein a first time series of events is indicative that the sensor signal has increased and wherein a second time series of the at least two time series of events is indicative that the sensor signal has decreased, wherein the signal conversion module is configured and arranged to provide the at least two time series of events to an event-driven signal expansion module and/or to an artificial neuron population,

an event-driven signal expansion module comprising at least a first layer of event-driven spiking artificial neurons, wherein the signal expansion module is configured and arranged to receive the at least one time series of events from the signal conversion module and to generate from each time series of events a plurality of outgoing time series of discrete events by means of the first layer of neurons and provide the plurality of outgoing time series to an artificial neuron population,

the artificial neuron population comprising a plurality of artificial event-driven spiking neurons arranged in an event-driven spiking neural network, wherein the neuron population is configured and arranged to receive the time series of discrete events, wherein the neuron population is arranged to recognize the physiological condition of the person based on the received events, wherein the neuron population is configured to provide one or more processed time series of events to,

a conditional detection module arranged and configured to receive the events from the neuron population and to output a trigger signal, when the events received from the artificial neuron population indicate that the physiological signal comprises a feature indicative of the physiological condition.

2 . The event-driven spiking neural network system implemented on a chip according to claim 1 , wherein the first layer of neurons in the signal expansion module is arranged and configured in a feed-forward configuration only.

3 . The event-driven spiking neural network system implemented on a chip according to claim 2 , wherein the neuron population is arranged as a recurrent neural network, wherein the neurons in the neuron population are arranged in a second layer.

4 . The event-driven spiking neural network system implemented on a chip according to claim 1 , wherein the condition detection module comprises a third layer of artificial neurons, wherein the third layer is arranged and configured in a feed-forward configuration only, wherein the third layer is trained to detect the physiological condition from the one or more processed time series of events.

5 . The event-driven spiking neural network system implemented on a chip according to claim 1 , wherein the system comprises the plurality of sensors configured and arranged to detect physiological signals providing a plurality of sensors signals to the signal conversion module, wherein the plurality of sensors is configured and arranged to detect a physiological signal different from the physiological signals that the remaining sensors of the plurality of sensors are configured and arranged to detect.

6 . The event-driven spiking neural network system implemented on a chip according to claim 1 , wherein the system is configured to identify a plurality of physiological conditions from the at least one detected physiological signal.

7 . The event-driven spiking neural network system implemented on a chip according to claim 1 , wherein the neurons comprise an input bus for events, wherein the input bus is connected to an event filter comprised by each neuron, wherein the event filter is configured to update a neuron state of the neuron for selected events received from the input bus, wherein the neuron further comprises an event generator configured to generate an event, when the neuron state of the neuron exceeds or falls below a predefined threshold value.

8 . The event-driven spiking neural network system implemented on a chip according to claim 7 , wherein each neuron of the neuron population comprises a balancer configured and arranged to adjust an output frequency of events of the neuron based on a frequency of incoming events leading to an update of the neuron state and/or based on a cumulative effect of incoming events on the neuron state, leading to an update of the neuron state.

9 . The event-driven spiking neural network system implemented on a chip according to claim 1 , wherein the system is connected or connectable to an external computerized system, wherein the condition detection module is configured to provide the trigger signal to the external system.

10 . The event-driven spiking neural network system implemented on a chip according to claim 9 , wherein the system comprises a signal storage configured to store the sensor signal from the at least one sensor, wherein upon detection of a physiological condition, at least a portion of the sensor signal indicative of the physiological condition is provided to the external system.