Methods for neurological analysis and assessment based on clockwise trends of decision boundaries in learning tasks

Abstract

Methods in accordance with the present disclosure include administering a plurality of instances of a learning task to one or more subjects. The one or more subjects provide responses in accordance with the learning task and the responses are then graphed. Decision boundaries associated with the responses provided in each instance may differ between a first instance of the learning task and a second instance of the learning task. A trend is determined between a first decision boundary and a second decision boundary. Based on this trend, one or more of a neurological state, condition or impairment, a cognitive ability, or a cognitive flexibility may be assessed in connection with one or more subjects.

Description

METHODS FOR NEUROLOGICAL ANALYSIS AND ASSESSMENT BASED ON CLOCKWISE TRENDS OF DECISION BOUNDARIES IN LEARNING TASKSCross-Reference to Related Applications

[0001] This application claims the priority benefit, under 35 U.S.C. 119(e), of U.S. Application No. 63 / 593,209, filed October 25, 2023, and entitled “Methods for Neurological Analysis and Assessment Based on Clockwise Trends of Decision Boundaries in Learning Tasks,” which is incorporated herein by reference in its entirety.Background

[0002] Learning is a central component of adaptive behavior, facilitating the ability of organisms to navigate variable environments. Initially grounded in Pavlovian and operant conditioning frameworks (Pavlov, 1927; Skinner, 1938), the understanding of learning mechanisms has expanded through the integration of computational methodologies. However, many aspects remain partly unexplored, leaving critical questions unanswered.Summary

[0003] Current advancements in reinforcement learning algorithms provide deeper insights into the initial concepts, describing the mechanisms that govern agent-environment interactions (Sutton & Barto, 2018; LeCun et al., 2015). Temporal difference learning has fostered this progression, bridging computational theories with neurobiological processes and facilitating predictive learning (Gershman, 2017). The detection of dopaminergic reward prediction errors in the basal ganglia substantiates the merger of these theoretical concepts with biological observations (Schultz et al., 2017).

[0004] Building on this foundation, the field of neuroeconomics has emerged, representing the intersection of neuroscience, psychology, and economics, and focusing on detailed analysis of cost-benefit decision-making processes (O'Doherty, 2014). A central concern of this field is the investigation of approach-avoidance dynamics in daily decision-making processes, grounded in neural pathways (Rangel et al., 2008). Research of non-human primate, particularly involving macaques, has been critical, capitalizing on their neurological and behavioral similarities to humans and to study cognition and learning (Azarfar et al., 2018). This research has paved the way for instrumental electrophysiological studies,demonstrating the complexity of neural functioning during multi-faceted tasks and advancing our understanding across different species (Yuste et al., 2020).

[0005] Recent studies have increasingly focused on the orbitofrontal cortex due to its significant role in decision-making and learning processes, where it operates in response to variable reward scenarios (Schoenbaum et al., 2011; Rushworth et al., 2011). Data indicate that this region houses a valuation system essential to decision-making processes, tasked with evaluating potential outcomes against associated risks (Wilson et al., 2014). This research intersects with studies assessing the role of striatal networks in mediating approachavoidance behaviors, establishing a framework to understand the neurological underpinnings of these behaviors and their link to mood disorders, including depression (Amemori and Graybiel, 2012; Amemori et al., 2018).

[0006] Moreover, the exploration of primate decision-making within economic contexts has fostered a nuanced understanding of optimization strategies, integrating biological observations with economic theories for a comprehensive perspective (Padoa-Schioppa, 2011; Strait et al., 2016). This collaborative approach demonstrates the evolving landscape of modern neuroscience. Although the field has achieved substantial progress in delineating learning mechanisms through interdisciplinary research, significant areas remain unexplored. The integration of experimental, computational, and clinical research approaches has the potential to facilitate groundbreaking advancements in the diagnosis of mood disorders and the development of educational and therapeutic strategies (Dolan et al., 2018).

[0007] In conclusion, the field has progressed considerably in defining the neural substrates and computational models that underpin learning and decision-making processes. Non-human primate research emerges as a vital tool in this endeavor, offering not only a deeper understanding of known learning mechanisms but also the potential to uncover previously unidentified mechanisms.

[0008] The Inventors have recognized and appreciated that leveraging the genetic and neural similarities that these primates share with humans informs significant advances in analyzing and further understanding neural circuitry, which in turn facilitates improved methodologies for assessing cognitive ability and learning processes in primates more generally. More specifically, the Inventors have recognized and appreciated that experiments conducted with non-human primates and the data obtained therefrom inform novel and more effective educational strategies and therapeutic interventions grounded in a comprehensive understanding of biological learning processes.

[0009] In view of the foregoing, the present disclosure generally is directed to inventive methods for neurological analysis that facilitate, inter alia, assessments of neurological conditions or impairments, cognitive ability, and / or cognitive flexibility in a subject. As discussed in greater detail below, in some examples the disclosed methods involve administering various learning tasks to a subject (e.g., decision-making learning tasks, such as cost-benefit analysis decision making tasks) and calculating decision boundaries based on a subject’s performance of administered learning tasks over time. One or more trends in such decision boundaries (e.g., a change in slope of successive graphed decision boundaries, or a clockwise directional movement of successive decision boundaries) are employed to identify a functional state of one or more neurological structures, and / or assess / model a cognitive ability or cognitive flexibility of the subject.Brief Description of the Drawings

[0010] FIG. 1 illustrates an exemplary learning task in accordance with the present methodology.

[0011] FIG. 2 illustrates the behavior of two subjects during a first portion of a learning task.

[0012] FIG. 3 illustrates the behavioral progression of experimental subjects during application of a learning task.

[0013] FIG. 4 illustrates reward and punishment coefficients derived from GLM analysis for two subjects.

[0014] FIG. 5 A illustrates exemplary areas of a subject brain that are monitored during learning task events.

[0015] FIG. 5B illustrates neural activity spike counts around a variety of task events.

[0016] FIG. 5C illustrates statistically significant neural activity in two areas of a plurality of subject brains during learning task events.

[0017] FIG. 5D illustrates a stepwise regression of various predictors for firing rates observed during a cue / offer period of task events.

[0018] FIG. 6 illustrates a flowchart of an exemplary method in accordance with the present technology.

[0019] FIG. 7 illustrates a flowchart of an exemplary method in accordance with the present technology.

[0020] FIG. 8 illustrates a flowchart of an exemplary method in accordance with the present technology.

[0021] FIG. 9 illustrates a system in accordance with the present technology.Detailed Description

[0022] In this study, two primate subjects, designated as subjects P and D, were trained over the course of approximately two years, transitioning from a naive state to a level of expertise, in their use of directional saccades to report acceptance or rejection of a series of offers. These offers were indicated by simultaneous presentation of two cues, abutting horizontal bars whose variable lengths represented how much positive and negative reinforcement they would receive if they accepted the offer, represented by a juice incentive, and a deterrent, represented by an air-puff applied to the face. Subsequent to the initial training regimen, we introduced the subjects to a conventional approach-avoidance (Ap-Av) task, delineated in the upper segment of FIG. 1.

[0023] Experimental Task

[0024] The task initiation phase is characterized by a period of fixation, after which a composite cue manifests on the monitoring screen. This cue is visually bifurcated, with the red bar segment proportionally representing the volume or duration of the forthcoming first consequence (e.g., a reward), and the yellow bar segment indicating the intensity or duration of the impending second consequence (e.g., a punishment). During the ensuing Target / Decision phase, subjects are confronted with a binary choice: to "approach," thereby accepting the stipulated reward -punishment combination; or to "avoid," effectively declining the offer. Choosing the "approach" option triggers a jittered delay of between 600-800 ms before delivery of the first outcome (punishment), followed by a 1 s delayed delivery of the second outcome (reward).

[0025] Throughout the task, subjects are presented with a diverse array of approximately 200 reward-punishment combinations, furnished randomly, to closely mimic the complexity inherent in real-world decision-making processes, which routinely involve judicious deliberation of potential benefits against corresponding detriments.

[0026] During a first portion of the experiment, subjects may first fixate on a specific point for 1 second, shown as fixation 110 before encountering an offer constituted of a red and a yellow bar stacked on top of each other, shown as first cue 112. Following a 1.5-second cue presentation in first cue 112, they may be shown two peripheral targets in target / decision 114a: a cross and a square. The monkeys can accept or approach the offer by saccading to thecross, as seen in choice 114b, or reject or avoid it by saccading to the square, as seen in choice 114c. Accepting the offer leads to receiving a jittered delay of between 600-800 ms before delivery of an air puff determined by the size of the red bar, followed after 1 s by a reward gauged by the size of the yellow bar; rejecting it results in neither reward nor punishment. APavlovian trial 116 may be interspersed with the choice trials with a predetermined frequency, e.g., once every seven to eight trials. The Pavlovian trials may alternate between punishment and reward.

[0027] In a reversal task, a second fixation 120 is presented to the subject. A second set of cues 122 is then presented to the subject, with the cues reversed. In other words, where a first bar indicated a reward in first cue 112, the first bar will indicate punishment in second cue 122. Likewise, where a second bar indicated a punishment in first cue 112, the second bar will indicate reward in second cue 122. Target / decision 124a presents the subject with two peripheral targets, which the subjects may then accept / approach in choice 124b by saccading to the cross as with choice 114b, or reject or avoid it by saccading to the square in choice 124c as with choice 114c. A second section of Pavlovian cues 126 follows the same structure but with reversed associations for the red and yellow cues and likewise for the Pavlovian cues, red being linked with a punitive air puff and yellow with a juice reward.

[0028] Behavioral Findings

[0029] Upon mastery of the first portion, behavioral patterns like those depicted in FIG. 2 may emerge. The x-axis of plots 210 and 220 represents the percentage of reward, and the y- axis denotes the percentage of punishment. Trials in which subjects chose to approach or accept the offers may be marked with green crosses. Instances where subjects opted to avoid or decline the offers may be denoted by red squares. The black lines demarcate decisionboundaries, which may be mathematically derived through a linear regression model.

[0030] FIG. 2 depicts two plots, plot 210 and plot 220, illustrating the behavior of each subject during a first portion of an approach-avoidance learning task. The vertical y-axis represents the air puff amount (in percentage), and the horizontal x-axis denotes the reward amount (also in percentage). Choices made by the monkeys may be marked with green crosses for accepted or “approach” decisions, and red squares for rejected or “avoidance” decisions. Black lines demarcate the decision boundaries for each subject, highlighting their individual thresholds in choice making. Despite exhibiting different decision-making patterns, a general tendency is observed where both subjects may be more inclined to “approach” options characterized by higher rewards and smaller punishments, and to “avoid”those associated with greater punishments and lesser rewards. This figure encapsulates the nuanced decision-making patterns displayed by the monkeys when confronted with varying offers, elucidating their propensity to maximize rewards while minimizing punishments.

[0031] Upon the initiation of the task reversal of the second section, executed in distinct sessions, a crucial adaptive behavior was observed. In this segment, the red and yellow cues underwent a role reversal: the length of the red cue became indicative of the punishment (airpuff) magnitude or duration instead of the reward magnitude or duration, whereas the length of the yellow cue defined the reward's (juice) magnitude or duration instead of the punishment magnitude or duration. Confronted with this inversion, the subjects exhibited a notable adaptive strategy, not previously identified.

[0032] Their behavioral modifications transpired progressively, characterized by a trend of the decision boundaries until they realigned with their initial positions, a phenomenon documented in plots 310 and 320 shown in FIG. 3. The trend manifested as a change in slope between a first decision boundary and a second decision, a translation of a first decision boundary relative to a second decision boundary, and / or a change in angle of the first decision boundary relative to a horizaontal axis of a graph illustrating the first decision boundary compared to an angle of the second decision boudary relative to a horizontal axis of a graph illustrating the second decision boundary.

[0033] It is imperative to note the discernible discrepancy in the adaptation rates between the two subjects; while one completed the adaptive cycle within an approximate span of 3 months, the other necessitated around 7 months. This variation in learning rates, delineating one subject as a rapid learner and the other as a more gradual adapter, is congruent with patterns discerned in previous studies involving these individuals.

[0034] Generalized Linear Model

[0035] FIG. 4 illustrates plots 410 and 420 representing selection frequencies for rewards and punishments as a function of session as well as cue reversal. The analysis presented in FIG. 4 may delineate the sequential progression of reward and punishment coefficients for subjects P and D derived through a generalized linear model (GLM) analysis. This analytical approach is predicated on a binomial distribution and uses a logit link function to relate the conditional mean of the response variable to the predictors — in this instance, the normalized values of the red and yellow Pavlovian cue conditions. Data for the analysis were collected over a series of sessions, each marked along the x-axis, with the response variables — the rewardand punishment coefficients — plotted on the y-axis. Individual data points represent beta coefficients extracted from the GLM analysis, with reward and punishment responses denoted, respectively. Line 412 illustrates a coefficient for reward for a first subject, Monkey P; line 414 illustrates a coefficient for punishment for the first subject. Line 422 illustrates a coefficient for reward for a second subject, Monkey D; line 424 illustrates a coefficient for punishment for the second subject. These coefficients may be presented as percentage values, highlighting the probabilistic nature of the response variables.

[0036] Each session, represented as a point on the x-axis, exhibits the coefficients derived from the trials conducted during that session. The vertical dotted lines demarcate different stages (e.g., the reversal points) of the experiment, providing a clear representation of the transition between each phase, starting from 'B'. The zero reference line, indicated by a horizontal black line for each of plots 410 and 420, serves to distinguish between positive and negative coefficient values, providing a reference for the interpretation of the reward and punishment coefficients over the sessions.

[0037] The rigorous process of trial inclusion involved the elimination of trials with bar scores of zero and those falling under specified trial type categories, ensuring a dataset grounded in statistical relevance and facilitating a meaningful interpretation of the behavioral patterns over time. Moreover, the analysis accounted for correct and wrong responses, encapsulating them within a binary framework to furnish a dichotomous outcome variable for the GLM analysis.

[0038] In the initial sessions denoted as the baseline (B) period, it is evident that both subjects demonstrate a tendency towards positive beta coefficients in response to rewards (as represented by line 412 and line 422) and negative beta coefficients in reaction to punishments. With the introduction of cue reversals (highlighted by the vertical dotted black lines), there is a discernible period of adjustment characterized by a protracted reliance on previously established contingencies for determining responses. During this phase, there is a marked tendency to erroneously associate the yellow bar, now a reward indicator, with punishment, and conversely, the red bar with reward — a manifestation of mis-association grounded in earlier learned behavior. As exposure to the new contingencies continues, a gradual alignment to the correct associations can be observed, resulting in the eventual adherence to the appropriate beta coefficient signs, thereby showcasing the adaptive learning capabilities of the subjects involved.

[0039] This analytical depiction serves as a robust tool in the interpretation of behavioral dynamics over time, offering insights into the learning processes and adaptive behaviors exhibited by the subjects throughout the experimental timeline.

[0040] Neural Analysis

[0041] FIG. 5 A illustrates areas of an example subject brain 500Athat are monitored during learning task events. For example, probes (e.g., one or more probes 950 illustrated in FIG. 9) may be inserted into the pregenual anterior cingulate cortex (pACC 502) and / or the caudal orbitofrontal cortex (cOFC 504) or placed on a surface of the subject’s head to monitor neural activity in one, both, and / or other regions. In particular, these regions may be monitored during performance of one or more learning tasks in accordance with the present technology.

[0042] FIG. 5B illustrates neuronal responses to various task events through a series of 16 peri-event histograms, illustrated as histograms 510 through 546, each corresponding to a different event delineated in its respective panel title. The temporal dynamics may be captured within a window of 4 seconds, marked with the event occurrence at the 0-second point, thereby providing a symmetrical visualization of 2 seconds before and up to 2 seconds after the event. Each panel represents spike counts on the y-axis, organized in 50-millisecond bins to detail the neuronal activity in close proximity to the event times indicated on the x- axis.

[0043] Statistical validation of the observed differences in spike counts between a short interval following the event onset and a preceding period of equal length was carried out using chi-square tests, with a threshold of p < 0.05 denoting statistical significance. This choice of test is pertinent given its efficacy in handling categorical data — an apt representation of our binned spike counts — and aids in establishing whether any significant variations in spike responses may be evident in relation to the event onset, thereby allowing for a nuanced understanding of the underlying neuronal patterns and their potential shifts in response to task events.

[0044] FIG. 5C illustrates a plot 500C of neural activity of a plurality of subjects during an exemplary learning task in accordance with the present technology. This neural activity indicates a percentage of statistically significant neurons activating in cOFC 504 and pACC 502 in response to specific events for a learning task. This percentage may provide a measure of significant neural activity in a certain area during the event. Further, mapping neural activity in response to a stimulus such as a learning task stimulus may enable an analysis ofhow effective a particular rehabilitation strategy is, if a first stimulus generates higher or lower neural activity than a second stimulus, identifying effective learning strategies, or the like. Mapping and / or monitoring of neural activity as described herein may additionally enable a determination of a neuropsychiatric condition based on the neural activity, a determination of a measure of cognitive ability, a determination of cognitive flexibility, or the like.

[0045] FIG. 5D illustrates a number of neurons (referred to as “units”) activated in cOFC 504 that are correlated to categories of stimulus and / or reactions by the subject, including reward, aversion, expected utility, choice, reward choice, aversion choice, conflict, and reaction time. For example, the responses of 83 neurons in cOFC 504 are correlated to choice.

[0046] Multiple regression analyses were conducted to examine the patterns of neuronal activity during the cue period in the Approach-Avoidance (Ap-Av) reversal-only task, across all session types. To identify an optimal set of variables that linearly parameterized neuronal responses, we utilized a stepwise regression approach through Matlab's "stepwisefit" function (Mathworks™, Natick, MA). This method iteratively adds or removes variables from a linear model based on their statistical significance, as determined by a series of F-tests. It starts with a preliminary model and assesses the impact of including or excluding each variable on the model's explanatory power, using the P-value of the F-statistic as a measure for comparing models with or against each additional variable. The threshold for determining a variable's significance was a p-value less than 0.05.

[0047] In our analysis of the Ap-Av task, the selected variables included: reward value, airpuff duration, expected utility in the Ap-Av scenario, binary indicator for approach [1] or avoidance [0] choices, an interaction between choice and reward, an interaction between choice and airpuff duration, presence of decisional conflict, and reaction time. These variables were chosen for their potential to elucidate the underlying factors influencing neuronal activity patterns associated with decision-making in these tasks (e.g., as illustrated in FIG. 4).

[0048] Methods

[0049] Subjects

[0050] Two Macaca mulatta monkeys, one male (P, 15.4 kg) and one female (D, 9.1 kg), were utilized in experiments that strictly adhered to the Guide for Care and Use of Laboratory Animals of the United States National Institutes of Health. Furthermore, all experimentalprocedures were approved by the Committee on Animal Care of the Massachusetts Institute of Technology. Prior to training, both animals underwent a habituation process which acclimated them to a seated position within a monkey chair and ensured a proper facemask fit. For any surgical procedures, postoperative analgesics were administered to the monkeys. Prophylactic antibiotics were also injected intramuscularly on the day of the surgery and were continued daily for the subsequent week.

[0051] Task Procedures

[0052] The initial phase of training aimed to acquaint naive monkeys with the testing environment and basic task structure. Monkeys were first presented with a large central white fixation point occupying half of the screen. The goal was to have them fixate on this point, beginning with short durations, such as 50 milliseconds. Over time and with consistent training, the size of this fixation point was gradually reduced, and the duration of required fixation was increased. This progression continued until monkeys could consistently maintain their gaze for up to 1000 milliseconds.

[0053] After establishing consistent fixation behavior, the monkeys were introduced to the red and yellow bars. At this introductory stage, both bars represented positive outcomes. Different sizes of these bars were displayed to match different sized outcomes to emphasize the significance of size in the upcoming tasks and to instill the understanding that ‘size matters’.

[0054] As the training progressed, specific associations were assigned. The red bar continued to be associated with a positive outcome and was now consistently linked with a white cross. On the other hand, the yellow bar, which had initially indicated a positive outcome, was now consistently linked with a white square (undergoing a transformation later in the Ap-Av task, where it began to signify a negative outcome).

[0055] Training continued with the approach-approach (Ap-Ap) task, where both choices led to positive outcomes. During a typical choice trial, the sequence began with the monkey fixating on the central fixation point for up to 1000 ms. Then, the cue offer was presented, with two abutted red and yellow bars of random sizes, followed by peripheral symbols representing the outcomes linked to the red (cross) and yellow (square) bars, which appeared for 1 second. The monkey was required to fixate on one of these symbols for at least 200 ms, indicating its choice. A successful choice was confirmed with a yellow highlight circle around the selected symbol, whereas a failure to decide within the given time rendered the trial invalid.

[0056] Following a correct choice, an interval of 600-800 milliseconds would precede the outcome, corresponding to the chosen bar's association. If fixation was broken at any point during the task, a penalty, in the form of a 5 -second delay, would be imposed before the initiation of the next trial.

[0057] Upon mastering the Ap-Ap task, monkeys moved to the Ap-Av task, where the choices were associated with contrasting outcomes: positive for the red bar and negative for the yellow bar. In the Ap-Av task, the concept changes; instead of choosing between the two bars, the monkeys had to decide whether to approach or avoid the presented offers by saccading to the cross (to approach) or to the square (to avoid). Mixed into these primary tasks were the Pavlovian cue trials, which required no decision-making and constituted around 15% of all trials.

[0058] Several months into the rigorous training, a variation of the Ap-Av task was introduced. In this variation, the associations of the red and yellow bars were inverted, adding complexity and testing the monkeys' adaptability.Recording Setup

[0059] After the behavioral training phase, individualized recording chambers, informed by 3T MRI coordinates, were employed. These coordinates were derived from T1 -weighted (0.5 mm isotropic) and T2-weighted (0.35 mm isotropic) MRI scans, which provided a detailed visualization of the skull surface. Gray Matter Research fabricated these chambers, which were affixed onto both monkeys (P and D), covering a significant portion of the orbitofrontal cortex and the striatum.

[0060] Several weeks following the chamber installation, a craniotomy was carried out, uncovering the dura mater. Subsequently, the chamber was sealed using silicone elastomer (Kwik Sil). To pinpoint the exact coordinates for each electrode track, an MRI scan was executed with the chamber and grid filled with a solution of saline combined with a 5% gadolinium contrast agent.

[0061] Approximately two weeks’ post-craniotomy, the implant surgery took place.Chambers were fitted with grids that featured an array of holes (40 x 30), each with a diameter of 0.48 mm and a center-to-center distance of 1 mm. Probes were secured onto microdrives, facilitating screw-controlled adjustments (158 pm per complete turn). Using the MRI (T1 and T2-weighted images, 0.5 mm slice thickness) with the chamber and grid infusedwith saline and the 5% gadolinium contrast agent, the coordinates of every electrode track were accurately determined.

[0062] Ensuring all procedures were conducted under anesthesia and in sterile conditions, platinum-iridium electrodes (with impedances ranging between 0.8-1.5 MQ; FHC Inc., Bowdoin, ME) were implanted into the cortex post-recovery. Custom-made micromanipulators held all these electrodes in place on the grid. For Monkey P, 57 electrodes were implanted (20 in the pregenual anterior cingulate cortex and 37 in the caudal orbitofrontal cortex). Meanwhile, Monkey D was implanted with 48 electrodes, with 15 in the pregenual anterior cingulate cortex vicinity and 33 in the caudal regions of the orbitofrontal cortex proximity.

[0063] Econometric Modeling

[0064] When two options are linked to the cross and square targets, the likelihood of opting for the cross target can be characterized as pAP=l(l+exp(-(UAP-UAv))), where UAP and UAV represent the utility values associated with each option (Train, 2003). In the Ap-Av task, the relationship f=UAP - UAV wa best captured by a first-order linear model, identified by its minimal Bayesian information criterion (BIC) (Amemori and Graybiel, 2012).

[0065] For modeling purposes, the difference in utility UAP-UAV was parametrized as f(x,y)=ax+by+c. Here, x signifies the length of the red bar, y represents the length of the yellow bar, and the coefficients a, b and c, were established using generalized linear regression, tailored to the observed behavioral choices. Each utility was thus modeled as UAP=ax+by and UAV=C. The predicted outcome value, or the 'chosen value,' was computed as Ch V=PAP UAP + (1 -PAP) UAV.

[0066] Lastly, the cost-benefit ratio (CBR) — defined as the proportion of sensitivities to reward (benefit) versus sensitivities to airpuff (cost) — was represented as CBR= b / a. The decision boundary's slope, a / b acts as the inverse of the CBR. A change of slope of a first decision boundary relative to a second decision boundary may be determined in accordance with the present disclosure. In an embodiment, a change in slope of a second decision boundary relative to a first decision boundary may include a rotation of the second decision boundary relative to the first decision boundary. In an embodiment, a rotation of the second decision boundary relative to the first decision boundary may include a clockwise rotation.

[0067] Behavioral Model

[0068] To get a clearer understanding of effective adaptive strategies, we used linear advantage actor-critic agents, which fall under the umbrella of reinforcement learning algorithms. Their primary task was the approach / avoi dance challenge. At the heart of this challenge, the agent is presented with two observations, (xl, x2), representing perceived potential rewards and risks.

[0069] The agent s decision process utilizes the formula d l(xlrbl^r 2(x2rb2^ . Here, xl and x2 act as the monkey's perceptions of potential rewards and risks. The weights al and a2 define the relative importance the agent (or monkey) gives to each perception. Simultaneously, bl and b2 act as biases, adjusting the baseline values of the observations.

[0070] However, the agent doesn't directly act on d. Instead, it uses d to calculate the probability of choosing to approach using the formula papproach=l / l+e~d. This equation transforms the value of d into a probability between 0 and 1. The term inside the exponential function, e~d, serves as a non-linear transformation. When d is a large positive value, indicating a strong inclination towards approaching, Papproach gets closer to 1. Conversely, when d is a large negative value, signaling a strong inclination to avoid, Papproach approaches 0. In essence, papproach represents the agent's confidence level in its decision, based on its observations and internal biases.

[0071] With this probability in hand, the agent may decide to approach or to avoid. If it chooses to approach, it might either receive a reward or incur a cost, dependent on its initial observations. On the other hand, an avoidance decision results in no change. The task's complexity is amplified by the periodic reversals, wherein the roles of the observations unpredictably swap.

[0072] Impressively, this model demonstrates an ability to preempt unexpected reversals and mimic the behavioral patterns seen in monkeys. By adjusting the learning rates of the parameters, we could steer the direction of behavioral rotation, offering a deeper understanding of complex cognitive dynamics.

[0073] In sum, this model highlights the power of reinforcement learning in deciphering adaptive behaviors. The linear advantage actor-critic agents, through their calculated approach, managed to reflect the sophisticated behaviors typical of the natural world. This lays the groundwork for exploring how artificial agents can adapt to fluctuating environments, echoing the adaptability of animals.

[0074] A learning task in accordance with the disclosure may include an approach-approach task, an approach-avoidance task, a cost-benefit decision-making task, a benefit-benefit decision-making task, or any suitable learning task.

[0075] Discussion - Enabled Methodologies

[0076] The clockwise learning phenomenon observed in non-human primates during the modified Ap-Av task offers intriguing insights into cognitive adaptability and the underpinnings of learning. Several methodologies are enabled by the present disclosure:

[0077] Reveals Systematic Neural Processes: A predictable and systematic trend between a first decision boundary and a second decision boundary (i.e., clockwise rotation, translation, change of angle relative to a horizontal axis) suggests that there are structured neural processes underpinning decision-making and adaptation. Understanding these processes can shed light on the fundamental neural circuits and computational principles involved in learning. Analysis of a decision boundary of an approach-avoidance task in accordance with the disclosure may include identifying a functional state of a neurological structure in a brain of the subject based at least in part on a comparison of a first neuronal pattern and a second neuronal pattern.

[0078] Models of Cognitive Flexibility: The Inventors have recognized and appreciated that trends in decision boundaries implicate new mathematical or computational models of cognitive flexibility. Such models may offer insights into how brains generalize and transfer knowledge, as well as the mechanisms that allow for efficient unlearning and relearning. For example, a neurodegenerative task may be administered to a subject based on a comparison of a first neuronal pattern and a second neuronal pattern during administration of a first and second approach-avoidance task. An exemplary neurodegenerative task may include an approach-avoidance task, a reinforcement learning task, a visual task, a haptic task, an auditory task, or a read / write task; each neurodegenerative task may be designed to remove, degrade, or regrow alternative connections within a brain.

[0079] Basis for Comparing Neural Strategies: If such a clockwise learning strategy is unique to non-human primates, as we deem improbable, or is also observed in other species, including humans, it could serve as a basis for comparing neural strategies across species, leading to a more comprehensive understanding of evolutionarily conserved mechanisms or divergent cognitive strategies.

[0080] Potential for Rehabilitation and Treatment: Understanding structured learning patterns can inform rehabilitation strategies in humans. For example, if certain patterns ofadaptation are known, interventions for neurological disorders or injuries can be designed to leverage or counteract these patterns to optimize recovery. Rehabilitation and / or treatment may include stimulating an area of a brain based on a trend of a second decision boundary compared to a first decision boundary. Stimulating an area of a brain may include electrically stimulating an area of a brain using one or more electrodes, and / or providing a visual stimulus (e.g., one or more words, one or more lights, one or more images, one or more videos, one or more colors, one or more shapes, one or more patterns, etc), an audio stimulus (e.g., one or more sounds, one or more voices, one or more words, one or more tones, one or more songs, one or more noises, etc.), a haptic stimulus (e.g., one or more vibrations, one or more textures, one or more substances, one or more surfaces, one or more objects, one or more materials, etc.), a gustatory stimulus (one or more fluids [such as juice, water, etc.], one or more foods, one or more edible items, etc.), or an olfactory stimulus to a subject (e.g., one or more smells, one or more aromas, one or more odors, one or more perfumes, one or more colognes, one or more aerosols, etc.).

[0081] Insights into Individual Differences: Noting that the speed of transition varied among individual animals, this learning strategy provides a platform to delve deeper into individual differences in cognitive processing. Individual differences in cognitive processing may manifest as one or more changes to a measure of cognitive ability. A measure of cognitive ability may include a response time, a percentage of approaches or avoidances given an approach avoidance task, a length of time focusing on a single point on a screen, a level of neural activity, or the like. A method in accordance with the present technology may include determining a measure of cognitive ability based on a change of a slope of a second decision boundary relative to a slope of a first decision boundary. This can further our understanding of why certain individuals or animals may be more adaptable or resilient in the face of change.

[0082] Neurodevelopmental and Neuropsychiatric Implications: Understanding the normative patterns of learning and adaptation can help identify deviations from these patterns, which might be indicative of neurodevelopmental disorders or neuropsychiatric conditions. This could aid in early diagnosis and intervention.

[0083] Educational Implications: If such structured learning patterns can be translated to educational settings, it might offer insights into optimizing teaching methodologies, especially in situations where unlearning incorrect information is necessary. Unlearning incorrect information may include administering, to a subject, a neurodegenerative taskcorresponding to a structure of a subject’s brain associated with a storage of incorrect information.

[0084] Artificial Intelligence and Robotics: Insights from such systematic learning patterns could influence the design of algorithms in machine learning, artificial intelligence, and robotics, helping create systems that adapt in more human-like or biologically inspired ways. In an aspect, methods disclosed herein may include determining at least one machine learning (ML) parameter based at least in part on a trend of a second decision boundary relative to a first decision boundary. An ML parameter may include a class of ML model, a weight of an ML model, a training set of an ML model, a number of nodes of an ML model, a number of inputs and / or outputs of an ML model, or any suitable ML parameter. One or more weights of an ML model may be selected to correlate an output of an ML mode to a human output or output of another subject (such as a non -human primate).

[0085] In essence, the clockwise learning phenomenon that we have observed serves as a unique window into the mechanisms of neural adaptation. By diving deeper into this phenomenon, we can unlock a myriad of applications and insights spanning from basic neuroscience to practical interventions in medicine, education, and technology.

[0086] FIG. 6 is a flowchart of an example method 600 for neurological analysis, the method including blocks 610-670.

[0087] At block 610, method 600 includes administering, to a subject, a first session of a learning task, the first session of the learning task comprising one or more first cues, one or more second cues, one or more first consequences, and one or more second consequences. In some instances, the learning task may include a cost-benefit decision-making task, a benefitbenefit decision-making task, or an approach-avoidance decision-making task.

[0088] At block 620, method 600 includes identifying a first plurality of selections made by the subject and a second plurality of selections made by the subject.

[0089] At block 630, method 600 includes calculating a first decision boundary demarcating the first plurality of selections and the second plurality of selections using a linear regression model.

[0090] At block 640, method 600 includes administering a second session of the learning task to the subject, the second session of the learning task comprising one or more third cues, one or more fourth cues, one or more third consequences, and one or more fourth consequences.

[0091] At block 650, method 600 includes identifying a third plurality of selections made by the subject and a fourth plurality of selections made by the subject.

[0092] At block 660, method 600 includes calculating a second decision boundary demarcating the third plurality of selections and the fourth plurality of selections using the linear regression model.

[0093] At block 670, method 600 includes determining a trend of the second decision boundary relative to the first decision boundary.

[0094] Method 600 may optionally include stimulating an area of a brain based at least in part on the trend. The stimulating may include administering a rehabilitation strategy for the area of the brain, administering an additional learning task to the brain of the subject based at least in part on the trend, administering an additional learning task to the brain of a second subject based at least in part on the trend, and / or applying an educational stimulus at a first rate to the subject based at least in part on the trend. Method 600 may optionally further include identifying a functional state of a neurological structure in a brain of the subject based at least in part on the trend, modeling a cognitive flexibility of the subject based at least in part on the trend, and / or determining a measure of cognitive ability of the subject based at least in part on the trend. The measure of cognitive ability may be a measure of cognitive flexibility. Method 600 may further include determining a neuropsychiatric condition based at least in part on the trend. The trend may include a rotation of the second decision boundary relative to the first decision boundary, and the rotation may include a clockwise rotation. Additionally or alternatively, the rotation may include a counterclockwise rotation. Method 600 may further include identifying at least one difference between the subject and a second subject based at least in part on the trend.

[0095] FIG. 7 is a flowchart of an example method 700 for neurological analysis, the method including blocks 710-760.

[0096] At block 710, method 700 includes administering a first session of a learning task to a first subject, the learning task comprising a first cue and a second cue. In some instances, the learning task may include a cost-benefit decision-making task, a benefit-benefit decisionmaking task, or an approach-avoidance decision-making task.

[0097] At block 720, method 700 includes measuring a first neuronal pattern of the first subject during the first session of the learning task.

[0098] At block 730, method 700 includes reversing the first cue and the second cue of the learning task.

[0099] At block 740, method 700 includes administering a second session of the learning task to the first subject with the reversed first cue and the reversed second cue.

[0100] At block 750, method 700 includes measuring a second neuronal pattern of the first subject during the second session of the learning task.

[0101] At block 760, method 700 includes identifying a functional state of a neurological structure in a brain of the first subject based at least in part on the first neuronal pattern and the second neuronal pattern.

[0102] Method 700 may optionally include determining a measure of cognitive ability of the subject based at least in part on the first neuronal pattern and the second neuronal pattern, a neuropsychiatric condition based at least in part on the first neuronal pattern and the second neuronal pattern, or at least one difference between the first subject and a second subject based at least in part on the first neuronal pattern and the second neuronal pattern.

[0103] FIG. 8 is a flowchart of a method 800 for neurological analysis, the method including blocks 810-850.

[0104] At block 810, method 800 includes administering a first session of a learning task to a first subject, the learning task comprising a first cue and a second cue. In some instances, the learning task may include a cost-benefit decision-making task, a benefit-benefit decisionmaking task, or an approach-avoidance decision-making task.

[0105] At block 820, method 800 includes measuring a first neuronal pattern of the first subject during the first session of the learning task.

[0106] At block 830, method 800 includes reversing the first cue and the second cue of the learning task.

[0107] At block 840, method 800 includes administering a second session of the learning task to the subject with the reversed first cue and the reversed second cue.

[0108] At block 850, method 800 includes measuring a second neuronal pattern of the first subject during the second session of the learning task.

[0109] Method 800 may optionally include administering a rehabilitation strategy to the subject based at least in part on the first neuronal pattern and the second neuronal pattern. Administering a rehabilitation strategy may include administering an additional learning taskto a brain of the subject based at least in part on a comparison of the first neuronal pattern and the second neuronal pattern and / or applying an educational stimulus at a first rate to the subject based at least in part on a comparison of the first neuronal pattern and the second neuronal pattern. Method 800 may optionally include administering a rehabilitation strategy to a second subject based at least in part on the first neuronal pattern and the second neuronal pattern, identifying a functional state of a neurological structure in a brain of the subject based at least in part on a comparison of the first neuronal pattern and the second neuronal pattern, determining a measure of cognitive ability of the subject based at least in part on the first neuronal pattern and the second neuronal pattern, and / or determining a neuropsychiatric condition based at least in part on the first neuronal pattern and the second neuronal pattern.

[0110] FIG. 9 illustrates a system 900 for neurological analysis in accordance with the present technology. System 900 may include a computer 910 and a display 920. In some instances, computer 910 and display 920 may be combined into a single item, for example as a tablet computer. Display 920 may be used to present a task to subject 990 in accordance with the present technology, such as an approach-avoidance task.

[0111] A heart rate monitor 930, gaze monitoring camera 940, or the like may monitor physiological data of subject 990. Heart rate monitor 930 may provide an indication of a heart rate, heart rate variability, heart beat signals and / or patterns, or the like. This heart-related data may be correlated to portions of a task presented to subject 990, for example if a stress level of the subject increases upon reversal of an approach-avoidance task cue, the subject’s heart rate may increase and this heart rate increase may be detected by 930.

[0112] To determine a selection made by subject 990, system 900 may detect the subject’s gaze using gaze detection device 940. Gaze detection device 940 may include a camera that can calculate where subject 990 is looking based on pupil position relative to other key features of the subject’s face, such as their nose, their eyelids, eyebrows, edges of their eyes, or the like. Computer 910 or other suitable processing device may utilize determined gaze position to assess which target the subject has selected (e.g., approach vs. avoid).

[0113] One or more probes 950 may be used to identify and record neuronal activity during administration of one or more tasks. System 900 may include a single probe, two or more probes, four or more probes, ten or more probes, or any suitable number of probes. Probes 950 may include cranial implants, surface electrodes such as an electroencephalogram (EEG), or any suitable probes that allow for the monitoring of neuronal activity. Detected neuronal activity may be used by computer 910 or other suitable processing device to correlatereaction times and task decisions to particular neurons or areas of the brain to allow for a better understanding of how new information incorporated and learned by subjects as they react to a given task.

[0114] Reward device 960 and punishment device 970 are used to deliver corresponding rewards and punishments to the subject based on their responses to a given task. For example, reward device 960 may include a juice dispenser, speaker, display, computer, or any suitable device configured to provide a reward to subject 990. In aspects where reward device 960 includes a speaker, the speaker may be configured to emit a pleasing sound, tone, or other indication to subject 990 that their choice in a task was a positive choice. If reward device 960 includes a display, the display may be configured to present a positive image, video, or other indication to subject 990 that their choice in the task was a positive choice. Additionally or alternatively, a display of reward device 960 may indicate that based on the subject’s choice, they have received monetary compensation.

[0115] Punishment device 970 may include an fan or similar object configured to deliver a puff of air to a subject’s eye, which may result in very minor irritation that subject 990 finds negative. Punishment device 970 may further include a speaker, display, computer, or any suitable device configured to provide a punishment to subject 990. For example, a speaker of punishment device 970 may play an irritating sound to subject 990 if the subject selects an approach response when presented with a cue indicating a punishment. A display of punishment device 970 may indicate that a human subject has lost compensation due to their selection in a particular task. In an aspect, reward device 960 and punishment device 970 may be embodied as a single device, e.g., a single display configured to communicate an addition or subtraction of compensation to a human subject based on their response to a task.Conclusion

[0116] While various inventive embodiments have been described and illustrated herein, those of ordinary skill in the art will readily envision a variety of other means and / or structures for performing the function and / or obtaining the results and / or one or more of the advantages described herein, and each of such variations and / or modifications is deemed to be within the scope of the inventive embodiments described herein. More generally, those skilled in the art will readily appreciate that all parameters, dimensions, materials, and configurations described herein are meant to be exemplary and that the actual parameters, dimensions, materials, and / or configurations will depend upon the specific application or applications for which the inventive teachings is / are used. Those skilled in the art willrecognize, or be able to ascertain using no more than routine experimentation, many equivalents to the specific inventive embodiments described herein. It is, therefore, to be understood that the foregoing embodiments are presented by way of example only and that, within the scope of the appended claims and equivalents thereto, inventive embodiments may be practiced otherwise than as specifically described and claimed. Inventive embodiments of the present disclosure are directed to each individual feature, system, article, material, kit, and / or method described herein. In addition, any combination of two or more such features, systems, articles, materials, kits, and / or methods, if such features, systems, articles, materials, kits, and / or methods are not mutually inconsistent, is included within the inventive scope of the present disclosure.

[0117] Also, various inventive concepts may be embodied as one or more methods, of which an example has been provided. The acts performed as part of the method may be ordered in any suitable way. Accordingly, embodiments may be constructed in which acts are performed in an order different than illustrated, which may include performing some acts simultaneously, even though shown as sequential acts in illustrative embodiments.

[0118] All definitions, as defined and used herein, should be understood to control over dictionary definitions, definitions in documents incorporated by reference, and / or ordinary meanings of the defined terms.

[0119] The indefinite articles “a” and “an,” as used herein in the specification and in the claims, unless clearly indicated to the contrary, should be understood to mean “at least one.”

[0120] The phrase “and / or,” as used herein in the specification and in the claims, should be understood to mean “either or both” of the elements so conjoined, i.e., elements that are conjunctively present in some cases and disjunctively present in other cases. Multiple elements listed with “and / or” should be construed in the same fashion, i.e., “one or more” of the elements so conjoined. Other elements may optionally be present other than the elements specifically identified by the “and / or” clause, whether related or unrelated to those elements specifically identified. Thus, as a non-limiting example, a reference to “A and / or B”, when used in conjunction with open-ended language such as “comprising” can refer, in one embodiment, to A only (optionally including elements other than B); in another embodiment, to B only (optionally including elements other than A); in yet another embodiment, to both A and B (optionally including other elements); etc.

[0121] As used herein in the specification and in the claims, “or” should be understood to have the same meaning as “and / or” as defined above. For example, when separating items in a list, “or” or “and / or” shall be interpreted as being inclusive, i.e., the inclusion of at least one, but also including more than one, of a number or list of elements, and, optionally, additional unlisted items. Only terms clearly indicated to the contrary, such as “only one of’ or “exactly one of,” or, when used in the claims, “consisting of,” will refer to the inclusion of exactly one element of a number or list of elements. In general, the term “or” as used herein shall only be interpreted as indicating exclusive alternatives (i.e. “one or the other but not both”) when preceded by terms of exclusivity, such as “either,” “one of,” “only one of,” or “exactly one of.” “Consisting essentially of,” when used in the claims, shall have its ordinary meaning as used in the field of patent law.

[0122] As used herein in the specification and in the claims, the phrase “at least one,” in reference to a list of one or more elements, should be understood to mean at least one element selected from any one or more of the elements in the list of elements, but not necessarily including at least one of each and every element specifically listed within the list of elements and not excluding any combinations of elements in the list of elements. This definition also allows that elements may optionally be present other than the elements specifically identified within the list of elements to which the phrase “at least one” refers, whether related or unrelated to those elements specifically identified. Thus, as a non-limiting example, “at least one of A and B” (or, equivalently, “at least one of A or B,” or, equivalently “at least one of A and / or B”) can refer, in one embodiment, to at least one, optionally including more than one, A, with no B present (and optionally including elements other than B); in another embodiment, to at least one, optionally including more than one, B, with no A present (and optionally including elements other than A); in yet another embodiment, to at least one, optionally including more than one, A, and at least one, optionally including more than one, B (and optionally including other elements); etc.

[0123] In the claims, as well as in the specification above, all transitional phrases such as “comprising,” “including,” “carrying,” “having,” “containing,” “involving,” “holding,” “composed of,” and the like are to be understood to be open-ended, i.e., to mean including but not limited to. Only the transitional phrases “consisting of’ and “consisting essentially of’ shall be closed or semi-closed transitional phrases, respectively, as set forth in the United States Patent Office Manual of Patent Examining Procedures, Section 2111.03.

Claims

CLAIMSWhat is claimed is:

1. A method for neurological analysis, the method comprising: administering, to a subject, a first session of a learning task, the first session of the learning task comprising one or more first cues, one or more second cues, one or more first consequences, and one or more second consequences; identifying a first plurality of selections made by the subject and a second plurality of selections made by the subject; calculating a first decision boundary demarcating the first plurality of selections and the second plurality of selections using a linear regression model; administering a second session of the learning task to the subject, the second session of the learning task comprising one or more third cues, one or more fourth cues, one or more third consequences, and one or more fourth consequences; identifying a third plurality of selections made by the subject and a fourth plurality of selections made by the subject; calculating a second decision boundary demarcating the third plurality of selections and the fourth plurality of selections using the linear regression model; and determining a trend of the second decision boundary relative to the first decision boundary.

2. The method of claim 1, further comprising stimulating an area of a brain based at least in part on the trend.

3. The method of claim 2, wherein the stimulating comprises administering a rehabilitation strategy for the area of the brain.

4. The method of claim 2, wherein the stimulating comprises administering an additional learning task to the brain of the subject based at least in part on the trend.

5. The method of claim 2, wherein the stimulating comprises administering an additional learning task to the brain of a second subject based at least in part on the trend.

6. The method of claim 2, wherein the stimulating comprises: applying an educational stimulus at a first rate to the subject based at least in part on the trend.

7. The method of claim 1, further comprising: identifying a functional state of a neurological structure in a brain of the subject based at least in part on the trend.

8. The method of claim 1, further comprising: modeling a cognitive flexibility of the subject based at least in part on the trend.

9. The method of claim 1, further comprising determining a measure of cognitive ability of the subject based at least in part on the trend.

10. The method of claim 9, wherein the measure of cognitive ability is a measure of cognitive flexibility.

11. The method of claim 1, further comprising determining a neuropsychiatric condition based at least in part on the trend.

12. The method of claim 1, wherein the trend comprises a rotation of the second decision boundary relative to the first decision boundary.

13. The method of claim 12, wherein the rotation comprises a clockwise rotation.

14. The method of claim 12, wherein the trend comprises a translation of the second decision boundary relative to the first decision boundary.

15. The method of claim 1, further comprising identifying at least one difference between the subject and a second subject based at least in part on the trend.

16. The method of claim 1, wherein the learning task is a cost-benefit decision-making task.

17. The method of claim 1, wherein the learning task is a benefit-benefit decision-making task.

18. The method of claim 1, wherein the learning task is an approach-avoidance decisionmaking task.

19. A method for neurological analysis, the method comprising: administering a first session of a learning task to a first subject, the learning task comprising a first cue and a second cue; measuring a first neuronal pattern of the first subject during the first session of the learning task; reversing the first cue and the second cue of the learning task; administering a second session of the learning task to the first subject with the reversed first cue and the reversed second cue; measuring a second neuronal pattern of the first subject during the second session of the learning task; and identifying a functional state of a neurological structure in a brain of the first subject based at least in part on the first neuronal pattern and the second neuronal pattern.

20. The method of claim 19, further comprising:administering an additional learning task to the subject based at least in part on the functional state.

21. The method of claim 19, further comprising: comparing the functional state with a second functional state of a corresponding neurological structure in a brain of a second subject; and applying an educational stimulus at a first rate to the first subject based at least in part on the step of comparing and applying the educational stimulus at a second rate to the second subject based at least in part on the step of comparing.

22. The method of claim 19, further comprising determining a measure of cognitive ability of the subject based at least in part on the first neuronal pattern and the second neuronal pattern.

23. The method of claim 19, further comprising determining a neuropsychiatric condition based at least in part on the first neuronal pattern and the second neuronal pattern.

24. The method of claim 19, further comprising identifying at least one difference between the first subject and a second subject based at least in part on the first neuronal pattern and the second neuronal pattern.

25. The method of claim 19, wherein the learning task is a cost-benefit decision-making task.

26. The method of claim 19, wherein the learning task is a benefit-benefit decision-making task.

27. The method of claim 19, wherein the learning task is an approach-avoidance decisionmaking task.

28. A method for neural analysis, the method comprising:administering a first session of a learning task to a first subject, the learning task comprising a first cue and a second cue; measuring a first neuronal pattern of the first subject during the first session of the learning task; reversing the first cue and the second cue of the learning task; administering a second session of the learning task to the first subject with the reversed first cue and the reversed second cue; measuring a second neuronal pattern of the first subject during the second session of the learning task.

29. The method of claim 28, further comprising administering a rehabilitation strategy to the subject based at least in part on the first neuronal pattern and the second neuronal pattern.

30. The method of claim 29, wherein administering a rehabilitation strategy comprises administering an additional learning task to a brain of the subject based at least in part on a comparison of the first neuronal pattern and the second neuronal pattern.

31. The method of claim 29, wherein administering the rehabilitation strategy comprises: applying an educational stimulus at a first rate to the subject based at least in part on a comparison of the first neuronal pattern and the second neuronal pattern.

32. The method of claim 28, further comprising administering a rehabilitation strategy to a second subject based at least in part on the first neuronal pattern and the second neuronal pattern.

33. The method of claim 28, further comprising: identifying a functional state of a neurological structure in a brain of the subject based at least in part on a comparison of the first neuronal pattern and the second neuronal pattern.

34. The method of claim 28, further comprising determining a measure of cognitive ability of the subject based at least in part on the first neuronal pattern and the second neuronal pattern.

35. The method of claim 28, further comprising determining a neuropsychiatric condition based at least in part on the first neuronal pattern and the second neuronal pattern.

36. The method of claim 28, wherein the learning task is a cost-benefit decision-making task.

37. The method of claim 28, wherein the learning task is a benefit-benefit decision-making task.

38. The method of claim 28, wherein the learning task is an approach-avoidance decisionmaking task.

Images