Abstract: An interface system for a user is provided that includes an interface unit configured to engage and selectively apply a force on a lower jaw of the user. An actuating unit is configured to actuate the interface unit in order to selectively apply and adjust the force on the lower jaw. One or more sensors are configured to generate output signals conveying information related to sleep stages of the user during a sleep session. The interface system further includes a processor configured to receive the output signals from the one or more sensors and determine a current sleep stage of the user. The processor controls the actuating unit based on the current sleep stage to adjust the force applied by the interface unit on the lower jaw.

Abstract: An interface system for a user is provided that comprises an interface unit configured to engage and selectively apply a force on a lower jaw of the user. An actuating unit is configured to actuate the interface unit in order to selectively apply and adjust the force on the lower jaw. One or more sensors are configured to generate output signals conveying information related to sleep stages of the user during a sleep session. The interface system further comprises a processor configured to receive the output signals from the one or more sensors and determine a current sleep stage of the user. The processor controls the actuating unit based on the current sleep stage to adjust the force applied by the interface unit on the lower jaw.

Abstract: Described herein are systems and methods for training first and second neural network models. A system comprises a memory comprising instruction data representing a set of instructions and a processor configured to communicate with the memory and to execute the set of instructions. The set of instructions, when executed by the processor, cause the processor to set a weight in the second model based on a corresponding weight in the first model, train the second model on a first dataset, wherein the training comprises updating the weight in the second model and adjust the corresponding weight in the first model based on the updated weight in the second model.

Abstract: Typically, high NREM stage N3 sleep detection accuracy is achieved using a frontal electrode referenced to an electrode at a distant location on the head (e.g., the mastoid, or the earlobe). For comfort and design considerations it is more convenient to have active and reference electrodes closely positioned on the frontal region of the head. This configuration, however, significantly attenuates the signal, which degrades sleep stage detection (e.g., N3) performance. The present disclosure describes a deep neural network (DNN) based solution developed to detect sleep using frontal electrodes only. N3 detection is enhanced through post-processing of the soft DNN outputs. Detection of slow-waves and sleep micro-arousals is accomplished using frequency domain thresholds. Volume modulation uses a high-frequency/low-frequency spectral ratio extracted from the frontal signal.

Abstract: The present disclosure pertains to a system configured to facilitate prediction of a sleep stage and intervention preparation in advance of the sleep stage's occurrence. The system comprises sensors configured to be placed on a subject and to generate output signals conveying information related to brain activity of the subject; and processors configured to: determine a sample representing the output signals with respect to a first time period of a sleep session; provide the sample to a prediction model at a first time of the sleep session to predict a sleep stage of the subject occurring around a second time; determine intervention information based on the prediction of the sleep stage, the intervention information indicating one or more stimulator parameters related to periheral stimulation; and cause one or more stimulators to provide the intervention to the subject around the second time of the sleep session.

Abstract: The present disclosure pertains to a system and method for delivering sensory stimulation to a user during a sleep session. The system comprises one or more sensors, one or more sensory stimulators, and one or more hardware processors. The processor(s) are configured to: determine one or more brain activity parameters indicative of sleep depth in the user based on output signals from the sensors; cause a neural network to indicate sleep stages predicted to occur at future times for the user during the sleep session; cause the sensory stimulator(s) to provide the sensory stimulation to the user based on the predicted sleep stages over time during the sleep session, and cause the sensory stimulator(s) to modulate a timing and/or intensity of the sensory stimulation based on the one or more brain activity parameters and values output from one or more intermediate layers of the neural network.

Abstract: The invention provides a system for providing an audio stimulus to a sleeping user. The system includes a user sensor adapted to acquire user sleep data from the sleeping user and a processor, which is adapted to determine a sleep stage of the sleeping user based on the user sleep data. The processor is further adapted to determine a predicted influence of an audio stimulus on the sleep stage of the sleeping user and generate a control signal based on the predicted influence of the audio stimulus. The system further comprises an audio output device, adapted to receive the control signal and generate an audio stimulus based on the control signal.

Abstract: Typically, high NREM stage N3 sleep detection accuracy is achieved using a frontal electrode referenced to an electrode at a distant location on the head (e.g., the mastoid, or the earlobe). For comfort and design considerations it is more convenient to have active and reference electrodes closely positioned on the frontal region of the head. This configuration, however, significantly attenuates the signal, which degrades sleep stage detection (e.g., N3) performance. The present disclosure describes a deep neural network (DNN) based solution developed to detect sleep using frontal electrodes only. N3 detection is enhanced through post-processing of the soft DNN outputs. Detection of slow-waves and sleep micro-arousals is accomplished using frequency domain thresholds. Volume modulation uses a high-frequency/low-frequency spectral ratio extracted from the frontal signal.

Abstract: The present disclosure pertains to a system and method for delivering sensory stimulation to a user during a sleep session. The system comprises one or more sensors, one or more sensory stimulators, and one or more hardware processors. The processor(s) are configured to: determine one or more brain activity parameters indicative of sleep depth in the user based on output signals from the sensors; cause a neural network to indicate sleep stages predicted to occur at future times for the user during the sleep session; cause the sensory stimulator(s) to provide the sensory stimulation to the user based on the predicted sleep stages over time during the sleep session, and cause the sensory stimulator(s) to modulate a timing and/or intensity of the sensory stimulation based on the one or more brain activity parameters and values output from one or more intermediate layers of the neural network.

Abstract: The present disclosure pertains to a system configured to facilitate prediction of a sleep stage and intervention preparation in advance of the sleep stage's occurrence. The system comprises sensors configured to be placed on a subject and to generate output signals conveying information related to brain activity of the subject; and processors configured to: determine a sample representing the output signals with respect to a first time period of a sleep session; provide the sample to a prediction model at a first time of the sleep session to predict a sleep stage of the subject occurring around a second time; determine intervention information based on the prediction of the sleep stage, the intervention information indicating one or more stimulator parameters related to periheral stimulation; and cause one or more stimulators to provide the intervention to the subject around the second time of the sleep session.

Abstract: Described herein are systems and methods for training first and second neural network models. A system comprises a memory comprising instruction data representing a set of instructions and a processor configured to communicate with the memory and to execute the set of instructions. The set of instructions, when executed by the processor, cause the processor to set a weight in the second model based on a corresponding weight in the first model, train the second model on a first dataset, wherein the training comprises updating the weight in the second model and adjust the corresponding weight in the first model based on the updated weight in the second model.

Abstract: A system for training a neural network model, comprises a memory comprising instruction data representing a set of instructions and a processor configured to communicate with the memory and to execute the set of instructions. The set of instructions, when executed by the processor, cause the processor to acquire training data, the training data comprising: data, an annotation for the data as determined by a user and auxiliary data, the auxiliary data describing at least one location of interest in the data, as considered by the user when determining the annotation for the data. The set of instructions when executed by the processor, further cause the processor to train the model using the training data, by minimising an auxiliary loss function that compares the at least one location of interest to an output of one or more layers of the model and minimising a main loss function that compares the annotation for the data as determined by the user to an annotation produced by the model.