Systems and methods for audio recommendations for neurostimulation

Abstract

The system includes a memory and one or more processors. The memory stores weights for a machine learning model. The weights are trained on training data of a training set. The training data includes an input audio signal, patient attributes, and measured brain response signals. The one or more processors are configured to receive the audio signal, apply the audio signal to the machine learning model to determine a predicted brain response signal for the audio signal, and generate a flag for the audio signal based on whether the predicted brain response signal satisfies one or more stimulus criteria.

Description

[Technical Field]

[0001] (CROSS-REFERENCE TO RELATED APPLICATIONS) This application claims priority to U.S. Provisional Patent Application No. 63 / 433,535, filed December 19, 2022, the contents of which are incorporated by reference in their entirety.

[0002] FIELD OF THE DISCLOSURE The present disclosure relates generally to neurostimulation, including, but not limited to, systems and methods for audio recommendations for neurostimulation. [Background technology]

[0003] (background) Neural oscillations occur in humans and animals and involve rhythmic or repetitive neural activity within the central nervous system. Nervous tissue can generate oscillatory activity through mechanisms within individual neurons or through interactions between neurons. Oscillations can manifest as either periodic fluctuations in membrane potential or rhythmic patterns of action potentials, which can produce oscillatory activation of postsynaptic neurons. The synchronized activity of groups of neurons can produce macroscopic oscillations, which can be observed by sensing electric or magnetic fields within the brain using techniques such as electroencephalography (EEG), intracranial EEG (iEEG), also known as electrocorticography (ECoG), and magnetoencephalography (MEG). Summary of the Invention [Means for solving the problem]

[0004] (summary) According to the systems and methods described herein, neural stimulation can be provided via rhythmic light stimulation presented simultaneously with auditory stimulation through music. The combination of music and light stimulation can induce a neural oscillation effect or stimulation. The combined stimulation can modulate, control, or otherwise influence the frequency of neural oscillation, providing beneficial effects on one or more cognitive conditions, cognitive function, the immune system, or inflammation, while mitigating or preventing adverse consequences on cognitive conditions or cognitive function. For example, the systems and methods of the present technology can treat, prevent, prevent, or otherwise influence Alzheimer's disease, or other cognitive disorders such as Parkinson's disease, dementia, etc.

[0005] Music can have a wide variety of tones, tempos, beats, and the like. Therefore, different musical pieces can produce different brain responses. For example, musical pieces with a weak onset or fast tempo changes (such as classical music or music from a violin) can produce a weaker brain response. On the other hand, musical pieces with a strong onset and slow tempo changes can provide a strong brain response. In addition, different musical pieces (even those with a strong onset and slow tempo changes) can stimulate the brain at different frequencies.

[0006] In various embodiments, as described in more detail below, the systems and methods described herein may be configured to receive or identify a music library (including multiple songs or musical pieces) for a patient. The systems and methods described herein may apply one or more songs from the music library to a machine learning model trained to determine a predicted response. The systems and methods described herein may determine whether the predicted response for the song meets criteria (e.g., a brain response at a target frequency or frequency range and at or above a target amplitude). If the song meets the criteria, the system may generate a flag for the song in the music library identifying the song as acceptable for audio stimulation. On the other hand, if the song does not meet the criteria, the system may generate a flag for the song in the music library identifying the song as unacceptable and / or generate a recommendation for a new song. By applying specific songs to machine learning models as described herein, the systems and methods may increase the effectiveness of audio-based stimulation, thereby producing more effective treatment results for various cognitive disorders.

[0007] In various aspects, the present disclosure is directed to a system and method for audio recommendation for neural stimulation. A memory may store weights for a machine learning model. The weights may be trained on training data of a training set, where the training data may include an input audio signal, patient attributes, and measured brain response signals. One or more processors may be configured to receive an audio signal and determine a predicted brain response signal for the audio signal by applying the audio signal to the machine learning model. The one or more processors may be configured to generate a flag for the audio signal based on whether the predicted brain response signal satisfies one or more stimulation criteria.

[0008] In some embodiments, an audio signal is identified in a library associated with the patient, and a flag is used to determine whether the audio signal should be used to provide audio stimulation to the patient. In some embodiments, the flag identifies one or more frequencies having an amplitude above a threshold in a predicted brain response signal for the audio signal. In some embodiments, the one or more processors may be configured to identify an audio signal from the plurality of audio signals based on one or more frequencies identified via the flag for the audio signal that match target frequencies for the patient's audio stimulation. The one or more processors may be configured to generate a recommendation including the identification of an audio signal for the patient's audio stimulation. [Brief explanation of the drawings]

[0009] The accompanying drawings are not intended to be drawn to scale. Like reference numbers and symbols in the various drawings indicate like elements. For purposes of clarity, not every component may be labeled in every drawing.

[0010] [Figure 1] FIG. 1 is a diagram illustrating the frequencies selected by the Oscillation Selection Module (OSM) as they relate to a specific underlying musical stimulus, and the range of frequencies present within each frequency band, according to an exemplary implementation of the present disclosure.

[0011] [Figure 2] Figure 2 is a schematic diagram illustrating, on the left, magnetoencephalography (MEG) recordings of the human auditory cortex recorded while subjects listened to rhythmic auditory stimuli at two different tempos, and, on the right, highlighting some of the brain areas that exhibited this response.

[0012] [Figure 3] FIG. 3 is a block diagram of a system for providing neurological stimulation according to an exemplary implementation of the present disclosure.

[0013] [Figure 4] FIG. 4 is a diagram illustrating the operation of the system of FIG. 3 with resulting brain stimulation according to an exemplary implementation of the present disclosure.

[0014] [Figure 5] 5-6 are diagrams showing exemplary stimuli provided by the system of FIG. 3 using different songs, where Panel A compares the auditory rhythmic frequency (i.e., onset spectrum) of the music with the frequency of an auditory 40 Hz pulse train, and Panel B compares the visual frequency stimulated by the system with the frequency of a visual 40 Hz pulse train, according to an exemplary implementation of the present disclosure. [Figure 6] 5-6 are diagrams showing exemplary stimuli provided by the system of FIG. 3 using different songs, where Panel A compares the auditory rhythmic frequency (i.e., onset spectrum) of the music with the frequency of an auditory 40 Hz pulse train, and Panel B compares the visual frequency stimulated by the system with the frequency of a visual 40 Hz pulse train, according to an exemplary implementation of the present disclosure.

[0015] [Figure 7] FIG. 7 is a schematic diagram of an output device for delivering visual stimuli, according to an exemplary implementation of the present disclosure.

[0016] [Figure 8] FIG. 8 is a block diagram of an example system using supervised learning, according to an example implementation of the present disclosure.

[0017] [Figure 9] FIG. 9 is a block diagram of a simplified neural network model according to an exemplary implementation of the present disclosure.

[0018] [Figure 10] FIG. 10 is a block diagram of an exemplary computer system in accordance with an exemplary implementation of the present disclosure. DETAILED DESCRIPTION OF THE INVENTION

[0019] (Detailed explanation) Before turning to the figures which illustrate certain embodiments in detail, it is to be understood that the present disclosure is not limited to the details or methodology set forth in the description or illustrated in the figures. It is also to be understood that the terminology used herein is for the purpose of description only and should not be regarded as limiting.

[0020] Neural oscillations can be characterized by their frequency, amplitude, and phase. These signal properties can be observed from neural recordings using time-frequency analysis. For example, EEG can measure oscillatory activity among groups of neurons, and the measured oscillatory activity can be categorized into frequency bands as follows: delta activity corresponds to the 0.5-4 Hz frequency band, theta activity corresponds to the 4-8 Hz frequency band, alpha activity corresponds to the 8-13 Hz frequency band, beta activity corresponds to the 13-30 Hz frequency band, and gamma activity corresponds to frequencies above 30 Hz.

[0021] Neural oscillations in different frequency bands can be associated with cognitive states or cognitive functions, such as perception, behavior, attention, reward, learning, and memory. Depending on the cognitive state or cognitive function, neural oscillations in one or more frequency bands may be involved. Furthermore, neural oscillations in one or more frequency bands may have beneficial or detrimental effects on one or more cognitive states or functions.

[0022] Neural entrainment occurs when an external stimulus of a specific frequency or combination of frequencies is perceived by the brain and triggers neural activity in the brain, resulting in neuronal oscillation at a frequency related to the specific frequency of the external stimulus. Thus, neural entrainment can refer to synchronizing neural oscillation in the brain using an external stimulus so that neural oscillation occurs at a frequency corresponding to the specific frequency of the external stimulus. Neural entrainment can also refer to synchronizing neural oscillation in the brain using an external stimulus so that neural oscillation occurs at a frequency corresponding to a harmonics, subharmonics, integer ratios, and combinations of the specific frequency of the external stimulus. The specific neural oscillation frequencies that can be observed in response to a set of external stimulus frequencies are predicted by a model of neural oscillation and neural entrainment.

[0023] Cognitive functions such as learning and memory involve coordinated activity across distributed subcortical and cortical brain regions, including the hippocampus, cortical and subcortical association areas, sensory regions, and the prefrontal cortex. Across different brain regions, behaviorally relevant information is encoded, maintained, and retrieved through transient increases in power and synchronization between neural oscillations that reflect multiple frequencies of activity.

[0024] In particular, oscillatory neural activity within the theta and gamma frequency bands is associated with encoding, maintenance, and retrieval processes during short-term, working, and long-term memory. Evoked gamma activity is related to working memory, which increases scalp-recorded and intracranial gamma-band activity during working / memory maintenance. Increases in gamma power dramatically track the number of items maintained in working memory. Using electrocorticography (ECoG), one study found that increases in gamma power tracked working / memory load within the hippocampus and medial temporal lobe as participants maintained sequences of letters or faces in working memory. Finally, other evidence indicates that hippocampal gamma activity subserves episodic memory and involves distinct sub-gamma frequency bands corresponding to encoding and retrieval stages.

[0025] Theta oscillations (4–8 Hz) have been linked to working and episodic memory processes. Intracranial EEG (iEEG) recordings demonstrate that during working memory, theta oscillations gate on and off (i.e., an increase and sustained amplitude followed by a rapid decrease in amplitude) across the encoding, maintenance, and retrieval phases. Other studies have observed increases in scalp-recorded theta activity during working / memory maintenance. Several investigations have concluded that scalp-recorded theta activity emerging from frontal / midline electrodes was the most robust neural correlate of verbal working / memory maintenance. Frontal / midline theta activity also follows working / memory load, i.e., increases in power and sustains as a function of the number of items maintained in working memory.

[0026] Several studies have found that gamma frequency auditory / visual stimulation can improve dementia or Alzheimer's disease (AD)-related biomarkers and pathophysiology and may provide neuroprotection when administered during the early stages of disease progression.

[0027] Music entrains and drives neural activity within multiple frequency ranges, and musical stimuli themselves can entrain and drive oscillatory neural activity involved in learning, memory, and cognition. In various embodiments of the present solution, the systems and methods described herein may detect, determine, identify, or otherwise exploit the brain's natural delta, theta, and gamma frequency responses to music by providing music as the sole auditory stimulus in systems and methods for treating, preventing, prophylactically, or otherwise influencing Alzheimer's disease, dementia, and / or other neurological or cognitive conditions. In some embodiments, audio stimuli are combined with visual stimuli in the delta, theta, and / or gamma frequency bands carefully tuned to synchronize with the delta, theta, and / or gamma frequency bands of the brain's response to the audio stimuli for enhanced therapeutic benefits. In some embodiments, additional frequencies and frequency bands can be targeted for stimulation to treat, prevent, and / or prevent disorders such as Alzheimer's disease, dementia, and / or other neurological or cognitive conditions, or Parkinson's disease.

[0028] Musical rhythms are organized into well-structured frequency combinations. For example, musical rhythms entrain neural activity in delta and theta frequencies by directly stimulating the brain at these frequencies. The fundamental beat frequency may correspond to neural activity in the delta frequency band. Subdivisions of the beat typically correspond to neural activity in the theta frequency band. In addition, musical rhythms can drive activity in delta and theta frequencies that are not explicitly present in the rhythm because musical rhythms contain structured frequency combinations. The frequencies observed in brain activity can include harmonics, subharmonics, integer ratios, and combinations of frequencies present in musical rhythms, and are predicted by simulations of neural oscillation and neural entrainment.

[0029] Musical rhythms can drive gamma neural activity in the brain in a manner distinct from the entrainment of delta and theta activity. The amplitude of endogenous gamma neural oscillations is modulated such that amplitude peaks synchronize with musical events (see Figure 2). Amplitude modulation of gamma neural activity reflects phase / amplitude coupling to lower frequency (e.g., delta and theta) neural activity.

[0030] Phase / amplitude coupling (PAC) can be or include the statistical dependence between the amplitude of oscillations in one frequency band and the phase of oscillations in another frequency band. For example, in theta / gamma phase / amplitude coupling, the peak of gamma amplitude corresponds to a specific phase of entrained theta activity. Thus, gamma activity is driven by entrained theta and delta activity.

[0031] The systems and methods described herein can provide feedback-based audio and / or visual stimulation by activating the brain's natural delta, theta, and gamma responses to music in a manner that does not interfere with music enjoyment. Because enjoyment is crucial for patient tolerance and protocol completion, the systems and methods described herein can motivate patient compliance with treatment by avoiding the harsh and unpleasant sounds of added audio waves in the gamma frequency band.

[0032] In some embodiments, the systems and methods described herein may incorporate, produce, or otherwise provide visual stimuli in the delta, theta, and / or gamma frequency bands to enhance frequencies that are important in music enjoyment. Such solutions may increase the effectiveness of the stimuli because visual stimuli in the gamma band are less aversive than auditory stimuli in the gamma band. In some embodiments, gamma stimuli can be combined with delta and theta stimuli to create visual stimuli that mimic the brain's natural response to musical rhythms.

[0033] In the systems and methods described herein, gamma stimulation can be amplitude-modulated into theta and / or delta frequency oscillations through phase / amplitude coupling to mimic auditory processing and increase the effectiveness and range of neural stimulation. Furthermore, the specific stimulation frequency is determined by the musical stimulation, and thus the stimulation frequency provided by the present solution can be varied within a stimulation session, reducing the potential for neural adaptation and therefore increasing stimulation effectiveness. In some embodiments, the systems and methods described herein can combine music listening with delta, theta, and / or gamma frequency visual stimulation to generate audiovisual stimulation that is both engaging and effective for the patient. In some embodiments, additional frequency bands may be employed via both the audio and visual stimulation.

[0034] In some embodiments, the systems and methods described herein may output an improved set of stimuli that amplify the brain's natural delta, theta, and gamma responses to music in a manner that does not create neural interference between the brain's natural oscillatory response to music and the added oscillatory auditory stimuli within the same frequency band. Specifically, in some embodiments, the systems and methods described herein may use a neural entrainment simulation to determine the frequencies of the brain's natural delta, theta, and gamma responses to music. The system may then reinforce and amplify the natural response to music by delivering the same delta, theta, and / or gamma frequencies within the visual stimuli. The simulation may include delta / theta / gamma phase / amplitude coupling to closely mimic the brain's auditory response and amplify the effect. Thus, the visual stimuli may not interfere with or counteract the brain's natural oscillatory response to music. Rather, the visual stimuli may amplify the brain's natural oscillatory response to music.

[0035] The systems and methods described herein are directed to outputting stimulation that induces neural stimulation via rhythmic light stimulation presented simultaneously with musical stimulation. The combination of music and rhythmic light pulses can induce electroencephalographic effects or stimulation. The combined stimulation can modulate, control, or otherwise influence the frequency of neural oscillations, providing beneficial effects on one or more cognitive conditions, cognitive function, the immune system, or inflammation (or other conditions), while mitigating or preventing adverse consequences on cognitive conditions or cognitive function, maximizing the acceptability, treatment tolerability, and completion of treatment protocols. For example, the systems and methods of the present technology can treat, prevent, prevent, or otherwise influence Alzheimer's disease (or other cognitive diseases or disorders).

[0036] The frequency of neural oscillations observed in a patient can be influenced by or correspond to the frequency of the musical rhythm and rhythmic light pulses. Thus, the system and method of the present solution can induce neural entrainment by outputting multimodal stimuli, such as musical rhythms and light pulses emitted at frequencies determined by analysis of the musical rhythm. This combined multimodal stimulation can synchronize electrical activity among groups of neurons based on a frequency or frequencies entrained and driven by the musical rhythm. Neural entrainment can be observed based on the aggregate frequency of oscillations produced by synchronized electrical activity in ensembles of neurons throughout the brain.

[0037] In some embodiments, additional outputs from the system may also include one or more stimulation units for generating tactile, vibratory, thermal, and / or electrical transcutaneous stimulation. Such stimulation units may include mobile devices, smartwatches, gloves, or other devices that may vibrate. In some embodiments, the output device may include a stimulation unit for generating an electromagnetic field or current, such as an array of electromagnets or electrodes, to deliver transcranial stimulation.

[0038] Referring to FIG. 1, depicted is a schematic diagram illustrating frequencies selected by an oscillation selection module (OSM) according to an exemplary implementation of the present disclosure as they relate to a specific underlying musical stimulus, and the range of frequencies present in each frequency band. As shown in FIG. 1, the diagram may include a breakdown of four frequencies that may be selected by the systems and methods described herein and as they relate to the underlying music and the range of frequencies present. In some embodiments, the systems and methods described herein may select one or more harmonically related frequencies within the delta, theta, and lower gamma (30-50 Hz) frequency ranges. In some embodiments, the gamma amplitude is modulated by the theta frequency to simulate theta / gamma phase / amplitude coupling. Also, in some embodiments, the theta amplitude is modulated by one or more delta frequencies to simulate delta / theta phase / amplitude coupling. Collectively, the foregoing thereby simulate a delta / theta / gamma oscillation hierarchy within the auditory cortex.

[0039] Continuing with reference to FIG. 1, an exemplary protocol for visual stimulus frequencies produced by the present system within the gamma, theta, and delta frequency bands in accordance with certain aspects of the present disclosure is illustrated. Panel A shows the time-domain waveform of a musical stimulus over a four-beat time interval and its onset calculated during preprocessing. Panel B shows the delta / theta / gamma combined change in luminance provided by the systems and methods described herein, while Panel C shows the same change within each frequency band.

[0040] Figure 2 shows MEG recordings of the human auditory cortex recorded while the subject listened to two rhythms with different tempos. Panel A of Figure 2 is a time / frequency map of signal power changes associated with rhythmic stimuli presented every 390 ms (2.6 Hz), which shows a periodic pattern of signal increases and decreases in the gamma frequency band. Panel B shows the same measurements for rhythmic stimuli presented every 585 ms (1.7 Hz). In the auditory cortex, gamma is amplitude modulated by delta and theta, and this pattern is simulated by the systems and methods described herein.

[0041] Panel D of Figure 1 illustrates the stimulation produced in the frequency domain by the systems and methods described herein. Collectively, these figures illustrate that gamma oscillations are effectively stimulated by the output provided by the devices in the range of frequencies surrounding the primary frequency. These additional frequencies are called sidebands, and they are caused by the amplitude modulation of the devices and methods from theta and delta frequencies. Also, each song played by the systems and methods described herein results in a different selection of frequencies within the delta, theta, and gamma ranges. Thus, over the course of several songs played via the systems and methods described herein, the output stimulates many gamma frequencies.

[0042] The device thus simulates amplitude modulation of stimulation provided in the gamma frequency band by the phase of stimulation provided in the delta and theta frequency bands, which mimics the brain's natural gamma / delta / theta phase / amplitude coupling response, thereby increasing both the tolerability and efficacy of the treatment. As described above, panel D of Figure 1 shows that gamma oscillations are effectively stimulated within a range of frequencies (sidebands) surrounding the dominant frequency. These sidebands are caused by amplitude modulation from theta and delta frequencies provided by the systems and methods described herein.

[0043] Each song played by the system may also result in a different selection of frequencies within the delta, theta, and gamma ranges. Thus, over the course of a session, different gamma frequencies are stimulated. In contrast, some solutions may stimulate only one frequency, with the general outcome being neuroadaptation, which leads to a reduced neural response. In some embodiments of the system, varying the frequency may avoid neuroadaptation and promote a robust neural response.

[0044] 3 and 4, depicted are a block diagram of a system 300 for providing neurological stimulation and a schematic diagram showing the operation of the system 300 with the resulting brain stimulation, according to an exemplary implementation of the present disclosure. The system 300 may include an auditory analysis system (AAS) 302 configured to receive auditory input, filter the acoustic signal, detect the onset of an acoustic event (e.g., a musical note or a drum hit), and adjust the gain of the resulting signal. In some embodiments, the AAS 302 may include a filtering module, an onset detection module, and an optional gain control module for filtering the signal, detecting the onset of the acoustic event, and adjusting the gain of the resulting signal, respectively.

[0045] In some embodiments, the AAS 302 may be configured to preprocess the auditory stimulus, auditory input, or audio signal 304 to provide multi-channel rhythmic input (e.g., musical note onset). In some embodiments, as described below with reference to FIGS. 8 and 9, the AAS 302 may be configured to preprocess the audio signal 304 to determine whether the audio signal 304 is suitable for audio stimulus. In some embodiments, the auditory input or audio signal 304 is provided by the system, for example, by or through an integrated audio playback system that has access to a library of songs and / or other musical pieces. For example, the audio signal 304 may be a song from a music library included in a profile for the patient (e.g., managed by the profile manager 306). As another example, the audio signal 304 may be a song from a music library from a remote server or source (e.g., an internet-based audio player, playback device, etc.).

[0046] In some embodiments, system 300 may further comprise a user-accessible graphical display or input / output to enable a user (e.g., patient or therapist) to make selections from a library for playback. In other embodiments, in addition to or as an alternative to a built-in audio playback system, system 300 may include an auxiliary audio input to enable system 300 to receive input from a secondary playback system, such as a personal music playback device (e.g., an iPod®, MP3 player, smartphone, or equivalent). In some embodiments, in addition to or as an alternative to the auditory inputs described above, system 300 may include a microphone or similar means to enable system 300 to receive auditory input from ambient sounds, such as a live music performance or music broadcast from secondary speakers, such as a user's home stereo system. In embodiments in which audio signal 304 is received by the system through an auxiliary input, such as through a built-in playback system or an MP3 player, the system may further comprise headphones or integrated speakers to enable a listener to hear audio signal 304 in real time.

[0047] The system 300 may include a profile manager 306. The profile manager 306 may be or include a processor or internet-enabled software application that accesses non-transitory and / or random-access memory that stores data about one or more users or patients, such as identifying information (e.g., name or patient ID number) stored within information from previous therapies and / or a library of audio files (e.g., a music library), in addition to various user preferences, such as song selections. The profile manager 306 may be communicatively coupled to the AAS 302 to facilitate the selection, management, or otherwise control of auditory input or audio signals. In some embodiments, the AAS 302 may be configured to generate flags for audio signals 304 included within the music library of a profile for the patient. The AAS 302 may be configured to generate flags based on or in accordance with preprocessing of the audio signals, as described below with reference to FIGS. 8 and 9. The flags may identify the acceptability of the audio signals 304 for audio stimulation, the frequencies or frequency bands of the audio stimulation, the individual amplitudes of different frequencies, etc.

[0048] In some embodiments, the profile manager 306 may provide a user interface to prompt the user to select their personalized music preferences as auditory stimuli. Such an implementation can maximize the efficacy of a given system by stimulating the auditory and reward systems in patients suffering from dementia and early stages of cognitive decline.

[0049] The system 300 may include an entrainment simulator (ES) 308. The ES 308 may receive and process received audio signals (e.g., from the AAS 302) to simulate processing in the human brain. The ES 308 may simulate the processing of the audio signal and propose and output oscillatory signals to enhance the received audio signal, thereby increasing the therapeutic effect of the treatment. In some embodiments, the AAS 302 is operatively connected to the ES 308 and provides data to the ES 308 in the form of an initiation signal. In some embodiments, the ES 308 also interfaces with the profile manager 306, e.g., to recall patient data from previous therapy. In some embodiments, the ES 308 may simulate entrained neural oscillations and predict the frequency, phase, and amplitude of a human's neural response to music.

[0050] The ES308 may include one or more oscillatory neural networks designed to simulate neural entrainment. In embodiments, the artificial oscillatory neural network receives preprocessed auditory stimuli (music), entrains the simulated neural oscillations, and predicts the frequency, phase, and relative amplitude of a human neural response to the music. In some embodiments, the ES308 may include a deep neural network, an oscillator network, a set of mathematical formulas, an algorithm, or any other component configured to mimic an oscillatory neural network. The ES308 can be configured to predict the frequency, phase, and relative amplitude of oscillations in a typical human brain entrained and driven by any given musical stimulus. The ES308 can be configured to predict responses within at least the delta (1-4 Hz), theta (4-8 Hz), and low gamma (30-50 Hz) frequency bands.

[0051] The system 300 may include an oscillation selection module (OSM) 310. The OSM 310 may be communicatively coupled to the ES 308. The OSM 310 may receive input from the ES 308 and output one or more selected oscillation states as frequency, amplitude, and phase for the visual stimulus. The OSM 310 may be configured to select the most prominent oscillations within one or more predetermined frequency ranges (in a preferred embodiment, the delta, theta, and gamma frequency bands) for the visual stimulus. In some embodiments, the OSM 310 may couple visual gamma frequency stimuli to the beat and rhythmic structure of music through phase / amplitude coupling. The OSM 310 may select variable music-based frequencies within the delta, theta, and gamma ranges for visual stimuli to the user, which stimuli are produced by a brain rhythm stimulator, as described below.

[0052] The system 300 may include a brain rhythm stimulator (BRS) 312. The BRS 312 may be configured to generate, produce, or otherwise provide control signals for an output device 314 to provide audio and / or visual stimuli based on data from the OSM 310, the ES 308, and / or the AAS 302. The BRS 312 may be configured to use simulated neural oscillations and synchronize visual stimuli within a selected frequency range to musical rhythms via an output device 314, such as an LED light ring, as described below. In some embodiments, the BRS 312 may output rhythmic visual stimuli to a user. The BRS 312 may include a pattern buffer, a generation module, a regulation module, and filtering components, and may be operably connected to the output device 314, which may comprise a means for displaying rhythmic light stimuli. The BRS 314 may also interface with a profile manager 306, which stores data regarding one or more users or patients. Thus, in some embodiments, the information stored by the profile manager 306 may also include previously captured or user-selected preferences for other parameters, such as stimulation pattern, waveform, or color, preferred by the user / patient.

[0053] The output device 314 may include an LED light, a computer monitor, a TV monitor, goggles, a virtual reality headset, augmented reality glasses, smart glasses, or other suitable stimulation output device. In some embodiments, the output device 314 may be a stimulation unit for generating tactile, vibration, thermal, and / or electrical transcutaneous stimulation, such as in a wearable device, a smart watch, or a mobile device. In some embodiments, the output device 314 may include a stimulation unit for generating an electromagnetic field or current, such as an array of electromagnets or electrodes, to deliver transcranial stimulation.

[0054] Collectively, the BRS may be configured to (1) read a patient's profile from the profile manager, (2) select a pattern based on the profile, (3) read one or more selected oscillatory signals and / or states from the ES / OSM, (4) generate a pattern, (5) adjust the pattern based on the profile, and (6) display or output rhythmic stimuli on an output device. In some embodiments, the pattern refers to a light pattern and the output device refers to a visual output device.

[0055] System 300 may include a brain oscillation monitor (BOM) 316. The BOM 316 may provide neural feedback that can be used to optimize the frequency, amplitude, and phase of visually presented oscillations to optimize the frequency, phase, and amplitude of oscillations in the brain. In some embodiments, the BOM 316 may provide feedback to system 300 (e.g., to ES 308) so that ES 308 may adjust parameters to optimize the phase of the outgoing oscillation signal. The BOM 316 may include, interface with, or otherwise communicate with electrodes, magnetometers, or other components arranged to sense brain activity, signal amplifiers, filtering components, and feedback interface components. In some embodiments, the BOM 316 may provide feedback to ES 308 in the form of EEG signals. The BOM 316 may be configured to identify the frequency, phase, and amplitude of brain oscillations entrained by the stimulation. The BOM 316 may be configured to sense electric or magnetic fields in the brain, amplify brain signals, filter the signals to identify specific neural frequencies, and provide input to the ES 308, as described above. The BOM 316 may be configured to sense electric or magnetic fields in the brain and may include electrodes connected to an electroencephalogram (EEG), intracranial EEG (iEEG), also known as electrocorticography (ECoG), magnetoencephalography (MEG), and other systems for sensing electric or magnetic fields.

[0056] AAS302, profile manager 306, ES308, OSM310, BRS312, and BOM316 may each be or include any hardware, including a processor, circuitry, or any other processing component, including any of the hardware or components described below with reference to FIG. 10.

[0057] Collectively, system 300 may be configured to (1) receive auditory input, (2) simulate neural entrainment to a preprocessed auditory signal using one or more tuning simulators 308, which may include a multi-frequency artificial neural oscillator network, (3) couple oscillations in the network using phase / amplitude or phase / phase coupling, (4) use an adaptive learning algorithm to adjust coupling parameters and / or intrinsic parameters, and / or (5) select the most prominent oscillations within one or more frequency bands for presentation as visual stimuli via BRS 312, as described below.

[0058] In embodiments, the rhythmic visual stimuli selected for output to the user (as described below) may include delta, theta, and / or gamma frequencies, as well as theta / gamma and / or delta / gamma phase / amplitude coupling, to enhance naturally occurring oscillatory responses to musical rhythms. Sensory cortices in the brain (e.g., primary visual and primary auditory cortices) are functionally connected to areas important for learning and memory, such as the hippocampus and medial and lateral prefrontal cortices. Thus, coupling complex rhythmic visual stimuli, including delta, theta, and gamma frequency visual stimuli, to musical rhythms can drive theta, gamma, and theta-gamma coupling in the brain, activating neural networks involved in learning, memory, and cognition. This, in turn, can drive learning and memory circuits involved in music.

[0059] Referring now to Figures 5 and 6, depicted are diagrams showing exemplary stimuli using different songs and visual stimuli according to an exemplary implementation of the present disclosure. Specifically, Figures 5 and 6 show a comparison between the auditory and visual stimuli provided by the systems and methods described herein and a 40 Hz pulse train. Figures 5 and 6 illustrate the various frequencies of audio and visual stimuli and the 40 Hz pulse train provided by both the systems and methods of the present disclosure. Figures 5 and 6 each illustrate stimuli provided by different songs. As can be seen, the 40 Hz pulse train provides both audio and visual stimuli at a single frequency, which can be easily contrasted with the wide range of frequencies at which the systems and methods described herein provide both audio and visual stimuli.

[0060] Referring now to FIG. 7 , depicted is one example of an output device 314 for providing visual stimuli. The output device 314 is provided via a visual stimulus ring 700 comprising an LED light 702 operably connected to the system 300, including the BRS 312. In some embodiments, the visual stimulus ring 700 is positioned in front of the participant, who is asked to focus on the center indicated by reference character 701. In some embodiments, the visual stimulus ring 700 is positioned at a distance appropriate to stimulate the retina at a specific visual angle. For example, the ring 700 may be positioned at a distance appropriate to stimulate the retina at a visual angle of 0-15 degrees, or 10-60 degrees, or 15-50 degrees, or 15-25 degrees, or 18-22 degrees, or 19-21 degrees. In some embodiments, the visual stimulus ring 700 may be positioned at a distance appropriate to stimulate the retina at a visual angle of 20 degrees, where the greatest density of rods is found in the retina.

[0061] Although illustrated as a stimulation ring 700, various other output devices 314 may also be used as part of the system 300, either in conjunction with or to complement the stimulation ring 700. For example, in some embodiments, the output device 314 may include a head-wearable device. The head-wearable device may include a display and / or one or more speakers of a speaker system. The head-wearable device may include augmented reality glasses, virtual reality goggles, etc. The display of the head-wearable device may render visual patterns to the user. For example, if the head-wearable device includes augmented reality glasses, the augmented reality glasses may augment the user's environment viewable through the glasses with visual patterns. As another example, if the head-wearable device includes virtual reality goggles (or other non-AR goggles), the goggles may display visual patterns on displays adjacent to the patient's eyes. In some embodiments, the display of the head-wearable device may display separate visual patterns on each of the patient's eyes at different angles to provide visual stimulation to the patient. The one or more speakers may include in-ear speakers or plug-in earphones for each of the patient's ears, headphones, a speaker system (e.g., locally on a head-wearable device), etc. The one or more speakers may be configured to render an audio signal 304 to provide audio stimuli to the patient.

[0062] In some embodiments, the output device 314 may include multiple output devices 314. For example, the output device 314 may include an audio output device 314 and a visual output device 314. The audio output device 314 may be configured to receive control signals from the BRS 312 to render the audio signal 304 to the patient as an audio stimulus. Similarly, the visual output device 314 may be configured to receive control signals from the BRS 312 to render a visual pattern to the patient as a visual stimulus. The audio output device 314 may be or include headphones, plug-in earphones, a speaker system, etc. The visual output device 314 may include the stimulus ring 700, a display device (e.g., a television, tablet, smartphone, or other display), a head-wearable device including a display, etc.

[0063] Therefore, in a method according to one embodiment of the present solution, the system may perform the following process.

[0064] (A) receiving auditory input;

[0065] (B) filtering the acoustic signal;

[0066] (C) Detecting the onset of an acoustic event;

[0067] (D) simulating neural entrainment to the preprocessed auditory signal using one or more multi-frequency neural oscillator networks;

[0068] (E) Coupling oscillations in a network using phase / amplitude or phase / phase coupling;

[0069] (F) using an adaptive learning algorithm to adjust binding parameters and / or intrinsic parameters;

[0070] (G) selecting for display the most prominent oscillations within the delta, theta, and / or gamma frequency bands;

[0071] (H) generating a light pattern; and

[0072] (I) Displaying a rhythmic light on a visual output device.

[0073] In some embodiments, prior to receiving audio input, the system may implement a process that prompts the user to select an audio input source and / or to make a selection from a library of songs or compositions stored by the system.

[0074] Self-selected music, i.e., music selected by the individual patient and familiar to them, may be more effective at engaging a larger network of brain activity in brain regions, including the hippocampus, auditory cortex, and frontal lobe regions important for long-term memory, compared to music selected by others or music unfamiliar to the patient. Thus, listening to familiar music may be more effective at driving brain activity in older adults, activating more brain areas. Importantly, familiar music may drive greater activation in the hippocampus, an area important for memory.

[0075] Music selected by the listener may be more likely to be liked and familiar to the listener than music selected by researchers, and may be more effective in engaging brain activity. In particular, self-selected music may increase activity within the dopaminergic reward system in addition to activating the auditory system in the default mode network and in the brain's prediction processes. Prolonged music listening may also increase brain functional connectivity from sensory cortices to the dopaminergic reward system, which is involved in various motivational behaviors.

[0076] Thus, in some embodiments, the auditory stimuli may include music self-selected by the patient, which has the practical impact of maximizing whole-brain engagement. The systems and methods described herein may facilitate receipt of music recordings from patients while they simultaneously view engaging audio-visual displays including delta, theta, and gamma frequency stimuli, further improving patient compliance with the disclosed treatment protocols.

[0077] In some embodiments, prior to generating and displaying the light pattern, the system 300 may prompt the user to select a profile from an input device and / or user interface integrated within or coupled to the system 300. The system 300 may perform one or more of the following processes: (G2) reading a patient's profile from the profile manager 306; (G3) selecting a light pattern based on the profile; (G4) reading one or more oscillator signals from the ES 308; (H) generating a light pattern; and (H2) adjusting the light pattern based on the profile.

[0078] In some embodiments, system 300 may also optimize the frequency, phase, and / or amplitude of the outgoing oscillating signal based on data received from BOM 316. Thus, system 300 may intermittently or continuously perform one or more of the following additional processes: (J) receiving input from BOM 316, (K) providing input to ES 308, (L) combining the input through phase / phase combining, and (M) using an adaptive learning algorithm to adjust combining parameters and / or intrinsic parameters to optimize the frequency, phase, and amplitude of the outgoing oscillating signal.

[0079] Thus, the systems and methods of the present solution may provide neural stimulation to a user through the presentation of rhythmic visual stimuli at least simultaneously, synchronously, and coordinated with musical stimuli.

[0080] For example, in some embodiments, system 300 may generate and display light patterns based on system self-selection or based on profile data stored for individual users to be displayed simultaneously with musical stimuli. In some embodiments, system 300 may perform one or more of the following additional processes:

[0081] (A) selecting one or more oscillations within the delta, theta, and / or gamma frequency bands;

[0082] (B) generating a light pattern using the selected one or more oscillations; and

[0083] (C) Displaying the light pattern on a visual output device 314.

[0084] The system 300 may also reference a user's profile and select a light pattern based on the profile. The system 300 may first prompt the user to select a profile from an input device and / or user interface integrated within or coupled to the system 300 and read the patient's profile from the profile manager 306 to determine an appropriate light pattern to display.

[0085] As described herein, in some embodiments, the AAS 302 may receive auditory input through a microphone or auxiliary audio input, filter the acoustic signal, detect the onset of an acoustic event (e.g., a musical note or a drum hit), and adjust the gain of the resulting signal.

[0086] As described herein, in some embodiments, the ES308 may receive auditory input from the AAS302 and use the input to simulate neural entrainment to the preprocessed auditory signal using one or more multi-frequency neural oscillator networks, couple oscillations in the networks using phase / amplitude or phase / phase coupling, adjust coupling parameters and / or intrinsic parameters using an adaptive learning algorithm, and select oscillations for display within a predetermined frequency range based on the retrieved profile. The ES308 may also receive input from the BOM316 and provide the input to one or more multi-frequency neural networks, combine neural inputs through phase / phase coupling, and adjust coupling parameters using an adaptive learning algorithm to optimize the amplitude and phase of the outgoing oscillatory signal.

[0087] As described herein, in some embodiments, the BRS312 may read a patient profile from the profile manager 306, select a light pattern based on the profile, read one or more oscillatory signals from the ES308, select at least one of a delta frequency, a theta frequency, a gamma frequency, and / or a combination of frequencies whose frequency, amplitude, and phase are determined by the ES308, generate a rhythmic light pattern based on the selected frequencies, adjust the light pattern based on the profile, and display rhythmic visual stimuli on an LED, computer monitor, TV monitor, or other suitable light output device directed toward the eye.

[0088] As a result of the systems and methods described herein, the system may sense electric or magnetic fields in the brain, amplify brain signals, filter signals, and identify specific neural frequencies. In some embodiments, the system then collects output from the user's brain based on the brain's reception of visual and audio stimuli and provides this feedback to the ES308, which further optimizes the visual and audio stimuli.

[0089] The present systems and methods can entrain and drive oscillatory neural activity involved in learning, memory, and cognition. By providing music as the sole auditory stimulus in addition to visual stimuli in the delta, theta, and / or gamma frequency bands, the present systems and methods can serve as a method for treating, preventing, prophylactically preventing, or otherwise influencing Alzheimer's disease and dementia.

[0090] Music can have a wide variety of tones, tempos, beats, and the like. Therefore, different musical pieces can produce different brain responses. For example, musical pieces with a weak onset or fast tempo changes (such as classical music or music from a violin) can produce a weaker brain response. On the other hand, musical pieces with a strong onset and slow tempo changes can provide a strong brain response. In addition, different musical pieces (even those with a strong onset and slow tempo changes) can stimulate the brain at different frequencies.

[0091] In various embodiments, as described in more detail below, the systems and methods described herein may be configured to receive or identify a music library (including multiple songs or musical pieces) for a patient. The systems and methods described herein may apply one or more songs from the music library to a machine learning model trained to determine a predicted response. The systems and methods described herein may determine whether the predicted response for the song meets criteria (e.g., a brain response at a target frequency or frequency range and at or above a target amplitude). If the song meets the criteria, the system may generate a flag for the song in the music library identifying the song as acceptable for audio stimulation. On the other hand, if the song does not meet the criteria, the system may generate a flag for the song in the music library identifying the song as unacceptable and / or generate a recommendation for a new song. By applying specific songs to machine learning models as described herein, the systems and methods may increase the effectiveness of audio-based stimulation, thereby producing more effective treatment results for various cognitive disorders.

[0092] 8 and 9, depicted are exemplary systems 800, 900 for machine learning or artificial intelligence. The systems 800, 900 may be incorporated into the system 300 (e.g., the AAS 302, the profile manager 306, the ES 308, etc.). The systems 800, 900 may be configured to generate a flag for the audio signal 304. The flag may indicate whether the music corresponding to the audio signal meets stimulation criteria. The stimulation criteria may be patient-specific (e.g., amplitude above or equal to a predetermined threshold for a patient-specific target frequency) or may be general stimulation criteria (e.g., amplitude above or equal to a predetermined threshold for various frequencies that may be used for different patients). The systems 800, 900 may be trained on a training set including data from a patient population. The patient population may be or include living patients (e.g., undergoing or previously treated), test patients, etc. The training set data may include audio signals corresponding to a wide variety of songs and measured brain responses. The training set may further include patient attributes, which may include, for example, the patient's age, the type or severity of cognitive disorder, hearing ability (e.g., full hearing, partial hearing loss, or complete hearing loss), the patient's medical condition, diagnostic data, heart rate, etc. The measured brain responses may include measured brain oscillations from the BOM 316, such as EEG signals or other feedback generated by the BOM 316.

[0093] As described in more detail below, the system 800, 900 may be configured to generate a predicted brain response for a given song. For example, the predicted brain response may include a predicted brain response at specific frequencies (e.g., a response to a particular combination of delta / theta / gamma frequencies at a particular amplitude) or a general predicted brain response (e.g., whether any of the delta, theta, or gamma frequencies in the predicted brain response are above a certain amplitude). The system 800, 900 may be configured to generate a flag or indicator for each song in the music library (e.g., in response to a song being applied to the system 800, 800). The flag or indicator may indicate, for example, the effectiveness of the song for audio stimulation at a specific frequency, whether the song is acceptable for treatment, etc. In some embodiments, in response to a song with a flag or indicator indicating low effectiveness and a song selected by the patient for audio stimulation, the systems and methods described herein may select or suggest / recommend a different song for audio stimulation.

[0094] Referring to FIG. 8 , a block diagram of an exemplary system using supervised learning is shown. In some embodiments, the system shown in FIG. 8 may be included, incorporated into, or otherwise used by the AAS 302 described above. In some embodiments, the system shown in FIG. 8 may be included, incorporated into, or otherwise used by the profile manager 306 and / or the ES 308. The system 300 may be configured to use supervised learning to generate a flag or other indicator that identifies the effectiveness of an audio stimulus for a particular audio signal. In some embodiments, the system 300 may be configured to use supervised learning to generate a flag specifically for a patient (e.g., to identify the effectiveness of an audio stimulus for an audio signal at a particular frequency for the patient). In some embodiments, the system 300 may be configured to identify a flag for a particular song and select, suggest, or recommend an alternative flag if the song does not meet the criteria. Supervised learning is a method of training a machine learning model given input / output pairs. An input / output pair is an input with an associated known output (e.g., an expected output).

[0095] The machine learning model 804 may be trained on known input / output pairs so that the machine learning model 804 can learn how to predict a known output given a known input. Once the machine learning model 804 has learned how to predict a known input / output pair, the machine learning model 804 can operate on unknown inputs and predict an output. The machine learning model 804 may be trained based on general data and / or granular data (e.g., patient-specific data based on previous stimuli and outcomes) so that the machine learning model 804 can be trained specifically for a particular patient.

[0096] Training inputs 802 and actual outputs 810 may be provided to the machine learning model 804. The training inputs 802 may include audio signals 304 previously used to stimulate various patients. The training inputs 802 may also include patient attributes such as cognitive impairment, age, heart rate, medications, diagnostic test results, patient history, etc. Thus, the training inputs 802 may include data (in some embodiments, patient attributes) corresponding to songs used for patient audio stimulation via the output device 314. The actual outputs 810 may include feedback from the BOM 316 (such as EEG data or other brain signals measured by the BOM 316).

[0097] Inputs 802 and actual outputs 810 may be received from the AAS 302, the ES 308, and / or the BOM 316 and stored in one or more data repositories. For example, the data repository may contain a data set including multiple data entries corresponding to past treatments. Each data entry may include, for example, patient attributes, audio signals 304 used for audio stimulation for the patient, and feedback data from the BOM 316. Thus, the machine learning model 804 may be trained to predict brain responses for a given audio signal based on the training inputs 802 and actual outputs 810 used to train the machine learning model 804. In some embodiments, the machine learning model 804 may be trained to predict brain responses for a given audio signal used for audio stimulation for patients with a particular set of attributes based on the training inputs 802 and actual outputs 810 used to train the machine learning model 804.

[0098] The system 300 may include one or more machine learning models 804. In one embodiment, a first machine learning model 804 may be trained to predict general effectiveness for a given audio signal. For example, the first machine learning model 804 may use a training input 802 of the audio signal 304 and predict a predicted feedback output 806 for the patient by applying a current state of the first machine learning model 804 to the training input 802. A comparator 808 may compare the predicted output 806 with an actual output 810 of the patient feedback to determine the amount of error or discrepancy. For example, the predicted EEG signal (e.g., predicted output 806) may be compared to an actual EEG signal (e.g., actual output 810) from the BOM 316.

[0099] In some embodiments, the second machine learning model 804 may be trained to make one or more predictions regarding efficacy for a particular patient and / or cognitive condition or disease based on the predicted output from the first machine learning model 804. For example, the second machine learning model 804 may use the training inputs 802 of patient attributes and feedback from the BOM 316 to predict efficacy outputs 806 for patients with similar patient attributes by applying the current state of the second machine learning model 804 to the training inputs 802. The comparator 808 may compare the predicted outputs 806 to the actual outputs 810 of the actual EEG signals from the BOM 316 at specific or target frequencies and determine the amount of error or discrepancy.

[0100] In some embodiments, a single machine learning model 804 may be trained to determine efficacy (e.g., general or patient-specific) based on audio signals, patient attributes, feedback, and / or other data received from the system 300. That is, a single machine learning model may be trained to predict an efficacy output 806 of an audio signal 304 using training inputs of the audio signal 304, patient attributes, and feedback from the BOM 316 by applying the current state of the machine learning model 804 to the training inputs 802. A comparator 808 may compare the predicted output 806 with the actual output 810 (e.g., the resulting EEG signal from the BOM 316) and determine the amount of error or discrepancy. The actual output 810 may be determined based on data received from the BOM 316.

[0101] During training, the error determined by the comparator 808 (represented by the error signal 812) may be used to adjust weights within the machine learning model 804 so that the machine learning model 804 changes (i.e., learns) over time. The machine learning model 804 may be trained using, for example, a back-propagation algorithm. The back-propagation algorithm operates by propagating the error signal 812. The error signal 812 may be calculated at each iteration (e.g., each pair of training input 802 and associated actual output 810), batch, and / or epoch and propagated through the algorithmic weights within the machine learning model 804 so that the algorithmic weights adapt based on the amount of error. The error is minimized using a loss function. Non-limiting examples of loss functions may include a squared error function, a root-mean-square error function, and / or a cross-entropy error function.

[0102] The weighting coefficients of the machine learning model 804 may be adjusted to reduce the amount of error, thereby minimizing (or otherwise converging on) the difference between the predicted output 806 and the actual output 810. The machine learning model 804 may be trained until the error determined in the comparator 808 is within a certain threshold (or a threshold number of batches, epochs, or iterations is reached). The trained machine learning model 804 and associated weighting coefficients may then be stored in memory 816 or other data repository (e.g., a database) so that the machine learning model 804 can be employed on unknown data (e.g., not the training inputs 802). Once trained and validated, the machine learning model 804 may be employed during testing (or an inference phase). During testing, the machine learning model 804 may incorporate unknown data (e.g., audio signals and / or patient attributes) and predict brain response data (e.g., predict EEG responses to audio signals and the like).

[0103] 9, a block diagram of a simplified neural network model 900 is shown. Similar to system 800, neural network 800 may be incorporated into system 300 to determine the effectiveness of audio signal 304 for audio stimulation. Neural network model 900 may include a stack of distinct layers (oriented vertically) that convert a variable number of inputs 902 taken by an input layer 904 into outputs 906 at an output layer 908.

[0104] The neural network model 900 may include several hidden layers 910 between the input layer 904 and the output layer 908. Each hidden layer has a distinct number of nodes (212, 914, and 916). In the neural network model 900, the first hidden layer 910-1 has a node 912, and the second hidden layer 910-2 has a node 914. The nodes 912 and 914 perform specific calculations and are interconnected to nodes in adjacent layers (e.g., node 912 in the first hidden layer 910-1 is connected to node 914 in the second hidden layer 910-2, and node 914 in the second hidden layer 910-2 is connected to node 916 in the output layer 908). Each of the nodes (212, 914, and 916) sums values ​​from adjacent nodes and applies an activation function, allowing the neural network model 900 to detect nonlinear patterns in the input 902. Each of the nodes (212, 914, and 916) is interconnected by weights 920-1, 920-2, 920-3, 920-4, 920-5, and 920-6 (collectively referred to as weights 920). The weights 920 are adjusted to adjust the strength of the nodes during training. Adjusting the strength of the nodes improves the neural network's ability to predict accurate outputs 906.

[0105] In some embodiments, the output 906 may be one or more numerical values. For example, the output 906 may be a vector of real numbers that are subsequently classified by any classifier. In one example, the real numbers may be input to a softmax classifier. A softmax classifier uses a softmax function, or a normalized exponential function, to convert the real inputs into a normalized probability distribution over the predicted output classes. For example, a softmax classifier may indicate the probability that the output is in class A, B, C, etc. Thus, a softmax classifier may be employed because of the classifier's ability to classify various classes. Other classifiers may also be used to perform other classifications. For example, a sigmoid function makes a binary decision for one class classification (i.e., the output may be classified using label A, or the output may not be classified using label A).

[0106] The system 300 may be configured to generate a flag based on predicted brain response data. For example, the system 300 may be configured to generate a flag based on whether the predicted amplitude at one or more specific frequencies meets a threshold criterion. The specific frequencies may be or include frequencies that are specific to the patient (e.g., target frequencies to be used for audio and / or visual stimulation for the patient). The specific frequencies may be or include frequencies that can be used for audio or visual stimulation for any given patient. The system 300 may be configured to generate a low efficacy flag in response to the predicted amplitude (e.g., at one or more frequencies) not meeting a threshold criterion (e.g., less than or equal to a threshold amplitude). The system 300 may be configured to generate a high efficacy flag in response to the predicted amplitude meeting a threshold criterion (e.g., greater than or equal to a threshold amplitude). In some embodiments, the flag may include additional data or information, such as patient attributes, frequencies with the best efficacy, frequencies with low efficacy, etc.

[0107] The system 300 may be configured to identify a flag generated by the system 800 for the audio signal 304 in response to the audio signal 304 being loaded into the system 300 for use in audio stimulation. The system 300 may be configured to determine whether the audio signal meets stimulation criteria for a patient based on a flag attached to or otherwise associated with the audio signal. The system 300 may be configured to determine stimulation criteria based on patient attributes (e.g., type of cognitive or neurological disorder, target frequency). The system 300 may be configured to determine whether the flag identifies high (or non-low) efficacy at the target frequency, either generally or specifically. If the flag for a particular song identifies low efficacy, the system 300 may be configured to select or recommend a different song for use in providing audio stimulation (e.g., by applying a filter based on high efficacy at the target frequency).

[0108] FIG. 10 depicts an exemplary block diagram of an exemplary computer system 1000. The computer system or computing device 1000 can include or be used to implement a data processing system or component thereof. The computing system 1000 includes at least one bus 1005 or other communication component for communicating information and at least one processor 1010 or processing circuitry coupled to the bus 1005 for processing information. The computing system 1000 can also include one or more processors 1010 or processing circuits coupled to the bus for processing information. The computing system 1000 also includes at least one main memory 1015, such as a random access memory (RAM) or other dynamic storage device, coupled to the bus 1005 for storing information and instructions to be executed by the processor 1010. The main memory 1015 can be used to store information during execution of instructions by the processor 1010. Computing system 1000 may further include at least one read-only memory (ROM) 1020 or other static storage device coupled to bus 1005 for storing static information and instructions for processor 1010. A storage device 1025, such as a solid-state device, magnetic disk, or optical disk, is coupled to bus 1005 and may provide persistent storage of information and instructions.

[0109] The computing system 1000 may be coupled via the bus 1005 to a display 1035, such as a liquid crystal display or an active matrix display, for displaying information to a user. An input device 1030, such as a keyboard or a voice interface, may be coupled to the bus 1005 for communicating information and commands to the processor 1010. The input device 1030 may include a touch screen display 1035. The input device 1030 may also include a cursor control device, such as a mouse, trackball, or cursor direction keys, for communicating directional information and command selections to the processor 1010 and for controlling cursor movement on the display 1035.

[0110] The processes, systems, and methods described herein can be implemented by computing system 1000 in response to processor 1010 executing sequences of instructions contained in main memory 1015. Such instructions can be read into main memory 1015 from another computer-readable medium, such as storage device 1025. Execution of sequences of instructions contained in main memory 1015 causes computing system 1000 to perform the illustrative processes described herein. One or more processors in a multiprocessing arrangement can also be employed to execute instructions contained in main memory 1015. Hardwired circuitry can be used in place of, or in combination with, software instructions, with the systems and methods described herein. The systems and methods described herein are not limited to any specific combination of hardware circuitry and software.

[0111] An exemplary computing system is illustrated in FIG. 10, although the subject matter, including the operations described herein, can be implemented in other types of digital electronic circuitry, or in computer software, firmware, or hardware, or in combinations of one or more of them, including the structures disclosed herein and their structural equivalents.

[0112] Although several illustrative implementations have been described herein, it should be apparent that the foregoing are presented by way of illustrative, not limiting, examples. In particular, while many of the examples presented herein involve specific combinations of method acts or system elements, those acts and those elements can be combined in other ways to accomplish the same purpose. Acts, elements, and features discussed in connection with one implementation are not intended to be excluded from similar roles in other or multiple implementations.

[0113] The hardware and data processing components used to implement the various processes, operations, illustrative logic, logic blocks, modules, and circuits described in connection with the embodiments disclosed herein may be implemented or performed using general-purpose single- or multi-chip processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs) or other programmable logic devices, discrete gate or transistor logic, discrete hardware components, or any combination thereof, designed to perform the functions described herein. A general-purpose processor may be a microprocessor, or any conventional processor, controller, microcontroller, or state machine. A processor may also be implemented as a combination of computing devices, such as a combination of a DSP and a microprocessor, multiple microprocessors, one or more microprocessors in conjunction with a DSP core, or any other such configuration. In some embodiments, particular processes and methods may be performed by circuitry that is specific to a given function. The memory (e.g., memory, memory unit, storage device, etc.) may include one or more devices (e.g., RAM, ROM, flash memory, hard disk storage, etc.) for storing data and / or computer code for completing or facilitating the various processes, layers, and modules described in this disclosure. The memory may be or include volatile or non-volatile memory and may include database components, object code components, script components, or any other type of information structure for supporting the various activities and information structures described in this disclosure. According to an exemplary embodiment, the memory is communicatively coupled to the processor via a processing circuit and includes computer code for executing (e.g., by the processing circuit and / or processor) one or more processes described herein.

[0114] The present disclosure contemplates methods, systems, and program products on any machine-readable medium for performing various operations. Embodiments of the present disclosure may be implemented using existing computer processors, or by specialized computer processors for suitable systems incorporated herein or for other purposes, or by wired systems. Embodiments within the scope of the present disclosure include program products comprising machine-readable media for carrying or having machine-executable instructions or data structures stored thereon. Such machine-readable media can be any available medium that can be accessed by a general-purpose or specialized computer or other machine with a processor. By way of example, such machine-readable media can be RAM, ROM, EPROM, EEPROM, or other optical disk storage, magnetic disk storage, or other magnetic storage devices, or any other medium that can be used to carry or store desired program code in the form of machine-executable instructions or data structures and that can be accessed by a general-purpose or specialized computer or other machine with a processor. Combinations of the above are also included within the scope of machine-readable media. Machine-executable instructions include, for example, instructions and data that cause a general-purpose computer, a specialized computer, or a specialized processing machine to perform a certain function or group of functions.

[0115] The phraseology and terminology used herein is for purposes of description and should not be regarded as limiting. The use herein of "including," "comprising," "having," "containing," "involving," "characterized by," "characterized in that," and variations thereof, is meant to encompass alternative implementations consisting of the subsequently listed items, equivalents thereof, and additional items, as well as the subsequently listed items only. In one implementation, the systems and methods described herein consist of one, each combination of more than one, or all of the described elements, acts, or components.

[0116] Any reference herein to system and method implementations, or elements, or acts, in the singular can also encompass implementations that include a plurality of those elements, and any reference herein to any implementation, element, or act in the plural can also encompass implementations that include only a single element. Reference in the singular or plural is not intended to limit the disclosed systems or methods, their components, acts, or elements to single or multiple configurations. Reference to any act or element that is based on any information, act, or element can include implementations in which the act or element is based, at least in part, on any information, act, or element.

[0117] Any implementation disclosed herein can be combined with any other implementations or embodiments, and reference to "an implementation," "some implementations," "one implementation," or the like is not necessarily mutually exclusive and is intended to indicate that a particular feature, structure, or characteristic described in connection with that implementation can be included in at least one implementation or embodiment. Such terms as used herein do not necessarily all refer to the same implementation. Any implementation can be combined, inclusively or exclusively, with any other implementation in any manner consistent with the aspects and implementations disclosed herein.

[0118] Where a reference sign follows a technical feature in a drawing, the detailed description, or any claim, the reference sign is included to improve the clarity of the drawing, the detailed description, and the claim. Thus, neither the reference sign nor its absence has any limiting effect on the scope of any claim element.

[0119] The systems and methods described herein may be embodied in other specific forms without departing from their characteristics. Reference to any term expressing degree includes a variation of + / -10% from the given measurement, unit, or range unless expressly indicated otherwise. Coupled elements may be electrically, mechanically, or physically coupled to each other directly or by means of intervening elements. The scope of the systems and methods described herein is therefore indicated by the appended claims, rather than the foregoing description, and changes that come within the meaning and range of equivalents of the claims are intended to be embraced therein.

[0120] The term "coupled" and variations thereof include the direct or indirect joining of two members to one another. Such a joining may be static (e.g., permanent or fixed) or movable (e.g., removable or releasable). Such a joining may be achieved using two members joined to or directly connected to one another, two members joined to one another using a separate intervening member and optional additional intermediate members joined to one another, or two members joined to one another using an intervening member that is integrally formed with one of the two members as a single, unitary body. When "coupled" or variations thereof are modified by additional terms (e.g., "directly coupled"), the general definition of "coupled" provided above is modified by the plain language meaning of the additional terms (e.g., "directly coupled" means the joining of two members without any separate intervening members), resulting in a definition narrower than the general definition of "coupled" provided above. Such coupling may be mechanical, electrical, or fluidic.

[0121] References of "or" can be construed as inclusive, such that any term described using "or" can refer to either one, more than one, or all of the described term. A reference to "at least one of 'A' and 'B'" can include "A" only, "B" only, and both "A" and "B." Such references in conjunction with "comprising" or other open-ended terminology can include additional items.

[0122] Modifications of the described elements and operations, such as variations in the size, dimensions, structure, shape, and proportions of various elements, parameter values, mounting arrangements, material use, color, orientation, etc., can occur without substantially departing from the teachings and advantages of the subject matter disclosed herein. For example, elements shown as integrally formed can be constructed from multiple parts or elements, the positions of elements can be reversed or otherwise varied, and the nature or number of discrete elements or positions can be altered or varied. Other substitutions, modifications, changes, and omissions can also be made in the design, operating conditions, and arrangements of the disclosed elements and operations without departing from the scope of the present disclosure.

Claims

1. 1. A system comprising: a memory for storing weights for a machine learning model, the weights being trained on training data of a training set, the training data including an input audio signal, patient attributes, and measured brain response signals; one or more processors Equipped with The one or more processors: receiving an audio signal; determining a predicted brain response signal for the audio signal by applying the audio signal to the machine learning model; and generating a flag for the audio signal based on whether the predicted brain response signal satisfies one or more stimulation criteria; A system configured to:

2. 10. The system of claim 1, wherein the audio signal is identified in a library associated with a patient, and the flag is used to determine whether the audio signal should be used to provide audio stimulation to the patient.

3. The system of claim 1 , wherein the flag identifies one or more frequencies in the predicted brain response signal for the audio signal that have an amplitude above a threshold.

4. The one or more processors: identifying an audio signal from a plurality of audio signals based on the one or more frequencies identified via the flag for the audio signal that match a target frequency for audio stimulation of a patient; generating a recommendation for the patient's audio stimulation, the recommendation including an identification of the audio signal; The system of claim 3 configured to:

Images