# US 2016/0249846 A1 · Systems and method for detecting, diagnosing, and/or treatment of disorders and conditions

> Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt

- Number: US 2016/0249846 A1
- Document ID: US20160249846A1
- Status: Published application
- Kind: A1
- Application number: 15/048,370
- Filed: 2016-02-19
- Published: 2016-09-01
- Earliest claimed: 2015-02-19
- Assignees: Khalifa University of Science and Technology
- Applicants: Masdar Institute Of Science And Technology
- Inventors: Jerald YOO; Muhammad ALTAF; Chen ZHANG
- Web page: https://patentgrep.com/patent/US20160249846A1

## CPC

- A61B 5/4094: {Diagnosing or monitoring seizure diseases, e.g. epilepsy}
- A61B 5/0531: Measuring skin impedance
- A61B 5/291: for electroencephalography [EEG]
- A61N 1/36025: {for treating a mental or cerebral condition}
- A61N 1/36031: {using physiological parameters for adjustment}
- G16H 50/20: for computer-aided diagnosis, e.g. based on medical expert systems
- A61B 5/6814: {Head}
- A61B 2560/0209: adapted for power saving
- A61B 2562/164: the sensor is mounted in or on a conformable substrate or carrier
- A61N 1/36064: {Epilepsy}
- A61B: DIAGNOSIS; SURGERY; IDENTIFICATION (usefulness limited to only animals A61D)
- A61N: ELECTROTHERAPY; MAGNETOTHERAPY; RADIATION THERAPY; ULTRASOUND THERAPY (measurement of bioelectric currents A61B; surgical instruments, devices or methods for transferring non-mechanical forms of energy to or from the body A61B18/00; anaesthetic apparatus in general A61M; incandescent lamps H01K; infrared radiators for heating H05B)
- G16H: HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA

## Abstract

A SoC includes an AFE to receive a plurality of differential input channels and generate digitized data corresponding to the channels, and a classification processor configured to receive the digitized data from the AFE. The AFE includes a Dual-Channel Chopper to perform channel multiplexing of two channels while simultaneously chopping the channels, a Dual Channel Charge Recycled-AFE having an Chopper-Stabilized Capacitive-Coupled IA including bias sampling capacitors that store bias values associated with the first and second channels to enable swapping between the channel, and a DC servo loop (DSL) having a reduced setting time based on a reduction in a resistance of the pseudo-PMOS in response to engaging a system reset. The classification processor includes a Frequency-Time Division Multiplexing (FTDM) Feature Extraction (FE) engine and a Dual-Detector Architecture (D2A) classification processor. The FTDM-FE includes a plurality of FIFOs configured to store, in parallel, the digitized data corresponding to the channels, a plurality of BPF banks storing BPF coefficients, and a single BPF to calculate outputs of one specific bank of the BFP banks for all of the channels. The D2A processor receives the output from the FTDM-FE and estimates a beginning and end of a seizure using two LSVMs optimized for only sensitivity and specificity, respectively.

## Claims

### Claim 1

A system on chip (SoC) with reduction power consumption and area comprising:
an analog front-end (AFE) configured to receive a plurality of differential input channels and generate digitized data corresponding to said plurality of differential input channels, said AFE comprising:
a Dual-Channel Chopper (DC-CHOP) configured to perform channel multiplexing of a first and a second differential input channel of said plurality of differential input channels while simultaneously chopping said first and said second differential input channel;
a Dual Channel Charge Recycled-AFE (DCCR-AFE) having a Chopper-Stabilized Capacitive-Coupled IA (CS-CCIA) including bias sampling capacitors that store bias values associated with said first and said second differential input channel to enable switching between said channel; and
a DC servo loop, wherein a setting time of said DC servo loop is reduced based on a reduction in a resistance of the pseudo-resistor (PMOS) of said DC servo loop in response to engaging a system reset (nRST); and
a classification processor configured to receive said digitized data from said AFE, said classification process comprising:
a Frequency-Time Division Multiplexing (FTDM) Feature Extraction (FE) engine comprising:
a plurality of FIFOs configured to store, in parallel, said digitized data corresponding to said plurality of differential input channels;
a plurality of band-pass filter (BPF) banks storing BPF coefficients; and
a single BPF to calculate outputs of one specific bank of said plurality of BFP banks for all of said plurality of differential input channels; and
a Dual-Detector Architecture (D2A) classification processor configured to receive said output data from said FTDM-FE and to approximate a beginning and an end of a seizure, said D2A classification processor comprising a first Linear Support-Vector Machine (LSVM) optimized for only sensitivity and a second LSVM optimized only specificity.

### Claim 2 (depends on 1)

The SoC of claim 1, further comprising:
skin-electrode impedance variation adaptation circuitry comprising stimulation signal circuitry to generate electrical stimulating pulses to be applied to a patient's skin through one or more electrodes, said stimulation signal circuitry configured to generate said electrical stimulating pulses based on said electrical stimulating pulses such that said stimulation signal circuitry automatically adapts a number of pulses with respect to skin-electrode impedance variation to ensure constant charge delivery.

### Claim 3

A channel sharing analog front-end (AFE) comprising:
said AFE configured to receive a plurality of differential input channels, said AFE comprising:
a Dual-Channel Chopper (DC-CHOP) configured to perform channel multiplexing of a first and a second differential input channel of said plurality of differential input channels while simultaneously chopping said first and said second differential input channel; and
a Dual Channel Charge Recycled-AFE (DCCR-AFE) having a Chopper-Stabilized Capacitive-Coupled IA (CS-CCIA) including bias sampling capacitors that store bias values associated with said first and said second differential input channel to enable swapping between said channel.

### Claim 4 (depends on 3)

The AFE of claim 3, further comprising a first and a second programmable gain amplifier (PGA) configured to amplify each of said output channels of said CS-CCIA.

### Claim 5 (depends on 3)

The AFE of claim 3, further comprising a DC servo loop, wherein a setting time of said DC servo loop is reduced based on a temporarily reduction in a resistance of the pseudo-resistor (PMOS) of said DC servo loop in response to engaging a system reset (nRST).

### Claim 6

A analog front-end (AFE) comprising:
said AFE configured to receive a plurality of differential input channels, said AFE comprising a DC servo loop (DSL), wherein a fast setting time of said DSL is reduced based on temporarily reduction in a resistance of the pseudo-resistor (PMOS) of said DSL in response to engaging a system reset (nRST) followed by disengaging nRST.

### Claim 7 (depends on 6)

The AFE of claim 6, wherein when DSL's pseudo-resistor (PMOS) gate is connected to a ground (GND) when said nRST is engaged.

### Claim 8 (depends on 6)

The AFE of claim 6, wherein said DSL's pseudo-resistor (PMOS) gate is coupled to VDD when said nRST is disengaged.

### Claim 9 (depends on 6)

The AFE of claim 6, wherein said DSL has a settling time of 0.5 second or less.

### Claim 10

Skin-electrode impedance variation adaptation circuitry comprising:
stimulation signal circuitry to generate electrical stimulating pulses to be applied to a patient's skin through one or more electrodes, said stimulation signal circuitry configured to generate said electrical stimulating pulses based on said electrical stimulating pulses such that said stimulation signal circuitry automatically adapts a number of pulses with respect to skin-electrode impedance variation to ensure constant charge delivery.

### Claim 11 (depends on 10)

The skin-electrode impedance variation adaptation circuitry of claim 10, wherein said stimulation signal circuitry is configured to monitor said naturally occurring RC time constant from said electrical stimulating pulses to determine information related to skin-electrode impedance.

### Claim 12 (depends on 10)

The skin-electrode impedance variation adaptation circuitry of claim 10, wherein said stimulation signal circuitry includes transcranial electrical stimulation (tES).

### Claim 13 (depends on 11)

The skin-electrode impedance variation adaptation circuitry of claim 11, wherein said skin-electrode impedance variation adaptation circuitry does not include current injecting circuitry for impedance monitoring.

### Claim 14

A Frequency-Time Multiplexing Frequency Extraction Engine comprising:
an analog front-end (AFE) configured to receive a plurality of differential input channels and generate digitized data corresponding to said plurality of differential input channels; and
a classification processor configured to receive said digitized data from said AFE, said classification process comprising a Frequency-Time Division Multiplexing (FTDM) Feature Extraction (FE) engine comprising:
a plurality of FIFOs configured to store, in parallel, said digitized data corresponding to said plurality of differential input channels;
a plurality of band-pass filter (BPF) banks storing BPF coefficients, wherein each BPF processed in a separate frequency; and
a single BPF to calculate outputs of one specific bank of said plurality of BFP banks for all of said plurality of differential input channels.

### Claim 15 (depends on 14)

The SoC of claim 14, wherein said single BFP comprises:
a data cache for storing said digitized data multiplexed from said plurality of FIFOs; and
at least one coefficients cache for storing said BFP coefficients multiplexed from said plurality of BFP banks.

### Claim 16 (depends on 15)

The SoC of claim 15, where said plurality of channels is equal to N channels, and wherein digitized input data corresponding to said N channels are stored in parallel in N number of multi-tap FIFOs of said FTDM-FE.

### Claim 17 (depends on 14)

The SoC of claim 14, wherein said FTDM-FE implements 2 s overlapped window with 1 s increment epoch windowing design.

### Claim 18

L-SVM with digital hysteresis comprising:
an analog front-end (AFE) configured to receive a plurality of differential input channels and generate digitized data corresponding to said plurality of differential input channels; and
a Dual-Detector Architecture (D2A) classification processor configured to receive said digitized data from said AFE and to approximate a beginning and an end of a seizure, said D2A classification processor comprising a first classifier optimized for only sensitivity and a second classifier optimized only specificity, wherein outputs of said first and said second classifiers are summed and a digital hysteresis is applied to reduce false positives.

### Claim 19 (depends on 18)

The SoC of claim 18, wherein said first and said second classifiers include a first Linear Support-Vector Machine (LSVM) and a second LSVM, respectively.

### Claim 20 (depends on 18)

The SoC of claim 18, wherein said digital hysteresis includes a Patient Specific Threshold (PST), and wherein said PST is established by supervised learning or using a self-learning algorithm.
