patent.fyi
Applications  /  US 2026/0355521
United States Patent Application Publication

US 2026/0355521 A1

Published
Nov 12, 2026
Filed
Sep 22, 2023
Earliest date on the face
Sep 26, 2022
Classification
A61N 1/36139
First published here
Nov 19, 2026

Generated automatically. Not reviewed by a person. Not legal advice or an opinion.

Section 1 · machine-generated

What if?

Questions about how individual parts of claim 1 might be changed. A question may describe something different from what the claim language recites, or a version of it; no view is offered on which.

1.1Optical recording of neural activity

What ifWhat if the system records neural activity through an optical fiber and a photodetector?
Claim language“from one or more electrodes”

An optical fiber delivers excitation light to tissue expressing a calcium indicator and carries fluorescence back to a filtered photodetector. A digitizer records fluorescence, and the feedback controller uses changes in fluorescence to select stimulation settings.

1.2Artifact prediction from pulse telemetry

What ifWhat if the processor predicts the artifact from stimulation current and pulse timing alone?
Claim language“detecting, based on the information identifying the brain activity”

A hardware timer captures pulse timestamps, and an output current monitor reports the delivered current. The processor scales a stored decay waveform by the measured current and subtracts the predicted transient from subsequent neural samples.

1.2Recording gated by pulse timing

What ifWhat if the recorder suspends acquisition during each pulse and its recovery interval, then measures activity between pulses?
Claim language“detrending the brain activity”

A pulse trigger gates the recording amplifier during stimulation and holds acquisition off for a configured recovery interval. The processor calculates band power from the samples acquired after the gate reopens and uses that measurement for feedback.

1.3Separate input for feedback estimates

What ifWhat if the estimator sends its output directly to the stimulation controller through a dedicated input?
Claim language“inserting the estimated brain activity”

The recorder stores acquired samples in a raw data buffer. The estimator calculates band power from its corrected samples and sends the resulting scalar and timestamp to the controller through a separate queue.

1.4Stimulation from a stored schedule

What ifWhat if a clock and a stored schedule determine the stimulation settings?
Claim language“based on the estimated brain activity”

A timer advances through a table of current amplitudes, pulse widths, and pulse repetition rates. At each scheduled transition, the controller loads the corresponding settings into the pulse generator.

Section 2 · machine-generated

Variations

Further arrangements of the same subject matter, described here as of the publication date at the foot of this page.

Other field of use

Evoked response control in cochlear implants

A cochlear electrode array delivers a pulse through one contact and records the resulting auditory nerve response through another contact. A processor fits a decaying exponential to the residual polarization tail in the first recording window and uses the fitted decay to correct samples in the next window. The processor inserts the corrected samples into the response buffer and measures peak-to-peak response amplitude. The controller compares that amplitude with a target band and adjusts the charge of a pulse delivered later in the second window.

Component substitution

Recursive least squares artifact estimation

A recursive least squares estimator replaces wavelet decomposition with a regression model driven by the recorded stimulation trigger and pulse amplitude. After amplifier recovery, the estimator fits a weighted sum of decaying exponentials to first-window samples selected outside the expected evoked response. The processor holds the fitted weights fixed through the second window, subtracts the projected artifact from incoming samples, and writes the corrected values to the controller buffer. The controller uses the buffered signal amplitude to set the current of a pulse delivered later in that window.

Reference electrode cancellation of recovery tails

A recording front end samples an active bipolar pair and a separate bipolar reference pair positioned to capture the stimulation transient with low sensitivity to the local neural signal. A processor estimates the coupling gain from the reference pair to the active pair by regression over first-window samples after amplifier recovery. In the second window, the processor multiplies the live reference samples by that gain, subtracts the result from the active samples, and inserts the corrected samples into the control buffer. The controller measures the corrected peak amplitude and uses it to set the pulse width of a stimulus delivered later in the window.

Parameter or scale variation

Eight channel processing in ten millisecond windows

Eight bipolar recording channels sample at 4,000 samples per second per channel and feed a double buffer with two 10 ms windows. For each channel, a processor fits an offset and the amplitude of a decay with a calibrated time constant to 32 first-window samples acquired after a 2 ms blanking interval. The processor extrapolates each fitted tail, subtracts it from the first 24 samples of the second window, and writes the corrected samples to the second-window buffer. The controller compares the median root-mean-square amplitude across channels with a threshold and schedules the stimulus at 8 ms into the second window when the amplitude exceeds the threshold.

Combination with existing systems

Integration with intracranial seizure detection

An intracranial seizure detector receives corrected electrode samples through the same ring buffer it uses for continuous monitoring. Following a stimulation pulse, an artifact processor fits the recovery tail in the first window and subtracts its extrapolation from the initial samples of the second window. The processor writes the corrected samples to their original time positions, and the detector updates a rolling line-length measure from the buffer. The stimulation controller delivers a pulse later in the second window when the updated measure crosses its trigger threshold and the programmed refractory interval has elapsed.

Automated or data-driven variant

Stimulation artifact prediction from recorded examples

A training system pairs first-window electrode samples and pulse settings with artifact reference traces captured in a saline fixture and scaled using electrode impedance measurements. A temporal convolutional model learns to predict the residual artifact at each sample position in the next window from the first-window trace, pulse settings, and measured impedance. During operation, the controller subtracts the predicted artifact from incoming samples and writes the corrected values to the second-window buffer. The controller measures rolling signal power from those values and uses it to select the amplitude of a pulse delivered later in the window.

Specific version of a part

Delivery in a later window

The controller writes the selected amplitude and pulse width to shadow registers during the second period. A separate timer event transfers those settings to the output registers and starts the pulse generator at the beginning of the third period.

Differential recording between adjacent contacts

Two adjacent implanted contacts feed a differential amplifier. An analog low-pass filter precedes a 2 kHz digitizer that acquires neural activity and the residual stimulation transient.

Wavelet reconstruction of slow drift

A discrete wavelet transform decomposes the post-pulse samples into detail coefficients and a coarse approximation. The processor extrapolates the slope of the reconstructed coarse approximation over the next short interval and subtracts that projected trend from the samples acquired there.

Timestamped replacement in a buffer

A circular buffer indexes the second period by sample timestamp. The processor writes each corrected estimate into the corresponding sample slot before feature extraction reads the buffer.

Amplitude selection using two thresholds

The controller computes band power from the corrected samples and compares it with upper and lower thresholds. It increments the programmed pulse amplitude by one step above the upper threshold, decrements it by one step below the lower threshold, and holds it between the thresholds.

Section 3 · quoted from USPTO records

Other work

Passages reproduced word for word from two earlier-published U.S. patent documents, placed beside the language of claim 1. No characterization is offered. Each passage is marked A or B for the document it comes from; ¶ numbers are the paragraph numbers in the USPTO’s published full text of that document.

Document AUS 2008/0051647 A1
Published
Feb 28, 2008
Filed
May 11, 2007
Class
A61N 1/04
Document BUS 2010/0069996 A1
Published
Mar 18, 2010
Filed
Sep 15, 2009
Class
A61N 1/00
Claim 1
Document A
Document B
1.1receiving, from one or more electrodes, information identifying brain activity with a stimulus trend artifact signal for a first time period;
Specifically, in this situation, the stimulus artifacts will typically be manifested as relatively large displacements of (i) the baseline of the nerve signal, and (ii) the nerve signal itself. These large stimulus artifact displacements can interfere with the relatively modest amplitudes of the nerve action potentials (NAPs) obtained by the detector's surface electrodes thereby undermining the usefulness of the detected signal (i.e., the trace).
A · US 2008/0051647 A1 · ¶ 10
Recordings of neuronal action potentials (NAPs) evoked by close-by applied electrical stimulation provide a signal mixture containing two major components: (1) the desired NAPs, and (2) the applied electrical stimulus.
B · US 2010/0069996 A1 · ¶ 3
1.2detecting, based on the information identifying the brain activity for the first time period, a trend in the stimulus trend artifact for the first time period and detrending the brain activity without the stimulus trend artifact signal for a second time period that is to occur after the first time period;
The controller/monitor 50 measures and analyzes the detected trace signal output from detector 40, determines the stimulus artifact, and then compares the amplitude of the stimulus artifact contaminating the detected NAP to a pre-defined limit. When the stimulus artifact is outside that pre-defined limit and is falling downward (signifying that the tissue is under-discharged), the time duration of the negative pulse is increased, whereby to minimize or substantially eliminate the stimulus artifact.
A · US 2008/0051647 A1 · ¶ 114
Another method records the response to a stimulus at sub-threshold level to measure a template for the stimulus artifact which is then scaled to supra-threshold levels and subtracted from the recorded signal mixture.
B · US 2010/0069996 A1 · ¶ 7
1.3inserting the estimated brain activity without the stimulus trend artifact signal into the information for a second time period;
1.4determining, based on the estimated brain activity for the second time period, a brain stimulus for the second time period; and
When the stimulus artifact is outside that pre-defined limit and is falling downward (signifying that the tissue is under-discharged), the time duration of the negative pulse is increased, whereby to minimize or substantially eliminate the stimulus artifact. When the stimulus artifact is outside the pre-defined limit and is rising upward (signifying that the tissue is over-discharged), the time duration of the negative pulse is decreased, whereby to minimize or substantially eliminate the stimulus artifact.
A · US 2008/0051647 A1 · ¶ 114
1.5causing the brain stimulus to be applied the second time period.
The previously-obtained optimum time duration of the negative pulse is then used as the initial time duration for the negative pulse of the next stimulation, and the foregoing steps are then repeated.
A · US 2008/0051647 A1 · ¶ 114
In the masker probe method, a second probe pulse is sent within the neuron's refractory time which allows a template for the stimulus artifact to be measured.
B · US 2010/0069996 A1 · ¶ 5

An empty cell means only that no passage was reproduced there.

Statements in these documents
This invention relates to methods and apparatus for electrically stimulating a nerve and for detecting the evoked nerve action potentials (NAPs), for both diagnostic and therapeutic purposes.
A · US 2008/0051647 A1 · ¶ 5
Nos. 5,284,153 and 5,284,154 is that the system is susceptible to contamination by stimulus artifacts.
A · US 2008/0051647 A1 · ¶ 7
The removal of the unintentional recorded stimulus waveform (also called the stimulus artifact) from this mixture has proven to be a difficult task.
B · US 2010/0069996 A1 · ¶ 3
Cite aspatent.fyi, US 2026/0355521 A1, claim 1, first published 2026-11-19T12:45:34-05:00, https://patent.fyi/US20260355521/, SHA-256 7923071e3fb322f0f67d3edacbafc3335e677ceb3a20ff75c492cd94ccaf32e1
First published
Permanent addresshttps://patent.fyi/US20260355521/
Content fingerprint (SHA-256)7923071e3fb322f0f67d3edacbafc3335e677ceb3a20ff75c492cd94ccaf32e1
Methodv2026-11-18.2 · automated, not reviewed by a person
Correctionscorrections@patent.fyi