Abstract: A brain-computer interface (BCI) includes a multichannel stimulator and a decoder. The multichannel stimulator is operatively connected to deliver stimulation pulses to a functional electrical stimulation (FES) device to control delivery of FES to an anatomical region. The decoder is operatively connected to receive at least one neural signal from at least one electrode operatively connected with a motor cortex. The decoder controls the multichannel stimulator based on the received at least one neural signal. The decoder comprises a computer programmed to process the received at least one neural signal using a deep neural network. The decoder may include a long short-term memory (LSTM) layer outputting to a convolutional layer in turn outputting to at least one fully connected neural network layer. The decoder may be updated by unsupervised updating. The decoder may be extended to include additional functions by transfer learning.

Abstract: A brain-computer interface (BCI) includes a multichannel stimulator and a decoder. The multichannel stimulator is operatively connected to deliver stimulation pulses to a functional electrical stimulation (FES) device to control delivery of FES to an anatomical region. The decoder is operatively connected to receive at least one neural signal from at least one electrode operatively connected with a motor cortex. The decoder controls the multichannel stimulator based on the received at least one neural signal. The decoder comprises a computer programmed to process the received at least one neural signal using a deep neural network. The decoder may include a long short-term memory (LSTM) layer outputting to a convolutional layer in turn outputting to at least one fully connected neural network layer. The decoder may be updated by unsupervised updating. The decoder may be extended to include additional functions by transfer learning.

Abstract: At least one electrical brain signal is received from a patient and is demultiplexed into an efferent motor intention signal and at least one afferent sensory signal (such as an afferent touch sense signal and/or an afferent proprioception signal). A functional electrical stimulation (FES) device is controlled to apply FES to control a paralyzed portion of the patient that is paralyzed due to a spinal cord injury of the patient. The controlling of the FES device is based on at least the efferent motor intention signal. A demultiplexed afferent touch sense signal may be used to control a haptic device. The afferent sensory signal(s) may be used to adjust the FES control.

Abstract: At least one electrical brain signal is received from a patient and is demultiplexed into an efferent motor intention signal and at least one afferent sensory signal (such as an afferent touch sense signal and/or an afferent proprioception signal). A functional electrical stimulation (FES) device is controlled to apply FES to control a paralyzed portion of the patient that is paralyzed due to a spinal cord injury of the patient. The controlling of the FES device is based on at least the efferent motor intention signal. A demultiplexed afferent touch sense signal may be used to control a haptic device. The afferent sensory signal(s) may be used to adjust the FES control.

Abstract: Approaches presented herein enable a device for trapping nanoparticles. More specifically, the device comprises a dielectric layer, an electrically insulating lid, a plurality of trapping electrodes, and electrical circuit connectors. The dielectric layer has an exposed surface, which is structured to form a set of recesses in the dielectric layer. The recesses are dimensioned so as to allow nanoparticles (e.g. biomolecules) to be trapped. The electrically insulating lid extends above the exposed surface of the dielectric layer. A flow path is defined between the lid and the exposed surface, such that a liquid can be introduced in the flow path. The trapping electrodes are arranged opposite the lid with respect to the exposed surface to face respective ones of the recesses. This arrangement defines pairs, such that each pair associates one of the trapping electrodes with a respective one of the recesses.

Abstract: A microfluidic device includes: a microchannel defining a flow path; a Brownian motor structure comprising two or more sorting channels having distinct ratchet topographies, the Brownian motor structure in fluid communication with the microchannel; and a filter extending transversely to the microchannel, the filter configured to filter particles, subject to sizes thereof, in a liquid advancing along the flow path, whereby smaller particles of the liquid can pass downstream of the filter in the flow path, and larger particles of the liquid are directed to the Brownian motor structure to be sorted out according to sizes thereof via the sorting channels.

Abstract: A microfluidic device includes: a microchannel defining a flow path; a Brownian motor structure comprising two or more sorting channels having distinct ratchet topographies, the Brownian motor structure in fluid communication with the microchannel; and a filter extending transversely to the microchannel, the filter configured to filter particles, subject to sizes thereof, in a liquid advancing along the flow path, whereby smaller particles of the liquid can pass downstream of the filter in the flow path, and larger particles of the liquid are directed to the Brownian motor structure to be sorted out according to sizes thereof via the sorting channels.

Abstract: A brain-computer interface (BCI) includes a multichannel stimulator and a decoder. The multichannel stimulator is operatively connected to deliver stimulation pulses to a functional electrical stimulation (FES) device to control delivery of FES to an anatomical region. The decoder is operatively connected to receive at least one neural signal from at least one electrode operatively connected with a motor cortex. The decoder controls the multichannel stimulator based on the received at least one neural signal. The decoder comprises a computer programmed to process the received at least one neural signal using a deep neural network. The decoder may include a long short-term memory (LSTM) layer outputting to a convolutional layer in turn outputting to at least one fully connected neural network layer. The decoder may be updated by unsupervised updating. The decoder may be extended to include additional functions by transfer learning.

Abstract: A detection device can include a cavity structure forming a Fabry-Perot optical microcavity, an electrostatic trap, and a Brownian motor. The Fabry-Perot optical microcavity has two mirrors extending on each side of a reference plane in a spacer region between the two mirrors. The mirrors are configured to vertically confine radiation in the spacer region, e.g., with respect to a first direction perpendicular to the reference plane. The electrostatic trap is arranged in the spacer region. The trap includes a pit and the cavity structure is generally configured to confine radiation in the pit, laterally (e.g., with respect to a second direction parallel to the reference plane). The Brownian motor structure extends in the spacer region along said reference plane. This structure is adapted to laterally load particles in the pit of the electrostatic trap by moving such particles along the structure, in operation.

Abstract: At least one electrical brain signal is received from a patient and is demultiplexed into an efferent motor intention signal and at least one afferent sensory signal (such as an afferent touch sense signal and/or an afferent proprioception signal). A functional electrical stimulation (FES) device is controlled to apply FES to control a paralyzed portion of the patient that is paralyzed due to a spinal cord injury of the patient. The controlling of the FES device is based on at least the efferent motor intention signal. A demultiplexed afferent touch sense signal may be used to control a haptic device. The afferent sensory signal(s) may be used to adjust the FES control.