WO2014066876A1 - Optimization of stochastic-resonance stimulation - Google Patents

Abstract

Systems and methods are disclosed to determine a subthreshold stimulation to be applied to mechanoreceptors. Systems and methods to apply this stimulation to mechanoreceptors are also disclosed. One system includes at least one bias-signal inputting mechanism configured to apply a subthreshold stimulation to mechanoreceptors, at least one bias-signal generator coupled to the at least one bias-signal inputting mechanism and configured to provide a driving signal having a temporal variation to drive the at least one bias-signal inputting mechanism, a controller for controlling the at least one bias signal generator and the at least one bias signal inputting mechanism, and a power source providing electrical energy to the controller and the at least one bias signal generator. The temporal variation can be used to counteract the effect of endogenous fractal fluctuations on stimulation efficacy.

Description

OPTIMIZATION OF STOCHASTIC-RESONANCE STIMULATION

CROSS-REFERENCE TO RELATED APPLICATIONS

[0001] This application claims benefit under 35 U.S.C. § 119(e) of U.S. Provisional Application No. 61/719,466 filed October 28, 2012, which is hereby incorporated by reference in its entirety.

FIELD OF THE INVENTION

[0002] The present invention relates to methods and apparatuses for optimizing stochastic resonance stimulation and, more particularly, to applying temporally dependent stimulation.

BACKGROUND OF THE INVENTION

[0003] Various devices are available for foot support and injury prevention. For example, passive orthoses and braces are described in the art (e.g. U.S. Pat. No. 6,692,454 to Townsend et al. and U.S. Pat. No. 6,676,618 to Andersen). These rigid or semi-rigid devices are typically directed toward supporting the foot or ankle to prevent injury, correct skeletal alignment problems, or adjust posture. In so doing, they may effect beneficial changes in balance and gait, but do so by providing passive mechanical support.

[0004] Wearable massaging apparatuses for feet are also known. For example, wearable foot massagers are described in the art (e.g. U.S. Pat. No. 5,835,899 to Reilly, U.S. Pat. No. 5,913,838 to Reilly, and U.S. Pat. No. 6,464,654 to Montgomery). Massagers typically employ mechanical means of introducing significant deflections (i.e. suprathreshold stimulation) of the skin and subcutaneous tissue, including muscles.

[0005] Moreover, wearable foot heaters are described in the art (e.g. U.S. Pat. No. 6,657,164 to Koch). These devices are typically directed toward pain relief, encouraging blood flow in skin, and maintaining thermal status of the foot, rather than to improving balance or gait. Heaters typically employ electrical resistance means to apply elevated temperatures directly to the skin of the foot.

[0006] Further, it has been found that the function of sensory cells in the human nervous system can be improved by inputting a noise signal to the sensory cell to effectively lower the threshold of the sensory cell. Lowering the sensory cell threshold decreases the level of outside stimulation needed to cause the sensory cell to respond (i.e. fire) because sensory cells are typically threshold-based units. Thus, the sensory cell will respond to outside stimulation at a lower level that would not result in a sensory cell response at normal cell threshold lev-

1

- i - els. U.S. Pat. Nos. 5,782,873 and 6,032,074 to Collins disclose a method and apparatus for improving the function of sensory cells by lowering the threshold of the sensory cells.

[0007] While these devices and methods fulfill their respective particular objectives and requirements, the aforementioned patents do not disclose a method, device, and system for optimizing neurological stimulation of body parts, such as the foot and/or ankle to improve human balance and gait and for preventing injury.

SUMMARY OF THE INVENTION

[0008] In one aspect of the invention, a system for neurological stimulation comprises at least one bias-signal inputting mechanism configured to apply a subthreshold stimulation to mechanoreceptors, at least one bias-signal generator coupled to the at least one bias-signal inputting mechanism and configured to provide a driving signal to drive the at least one bias- signal inputting mechanism, a controller for controlling the at least one bias signal generator and the at least one bias signal inputting mechanism, and a power source providing electrical energy to the controller and the at least one bias signal generator. The driving signal includes a temporal variation.

[0009] In another aspect of the invention, a method for neurological stimulation comprises the act of providing at least one bias-signal inputting mechanism configured to apply a subthreshold stimulation to mechanoreceptors, at least one bias-signal generator configured to provide a driving signal to drive the at least one bias-signal inputting mechanism, a controller for controlling the at least one bias signal generator and the at least one bias signal inputting mechanism, and a power source providing electrical energy to the controller and the at least one bias signal generator. The method also comprises activating the signal generator and supplying a bias signal to stimulate the mechanoreceptors, the bias signal having a temporal dependence based on a determined therapeutic need of an individual.

[0010] In yet another aspect of the invention, a method for developing a stimulation profile comprises the act of measuring at least one predetermined factor at a plurality of times, at least two of the times being temporally distinct, the at least one predetermined factor exhibiting temporal fluctuations. The method further comprises the act of formulating a temporal profile using at least two of the temporally-distinct times. The method further comprises developing the stimulation profile using the temporal profile, the stimulation profile, when applied to mechanoreceptors via at least one bias signal inputting mechanism, improving system stability. [0011] Additional aspects of the invention will be apparent to those of ordinary skill in the art in view of the detailed description of various embodiments, which is made with reference to the drawings, a brief description of which is provided below

BRIEF DESCRIPTION OF THE FIGURES

[0012] FIG. 1 depicts a system for applying neurological stimulation having a temporal variation according to one embodiment.

[0013] FIG. 2 depicts a method of determining a threshold sensory performance for an individual according to one embodiment.

[0014] FIG. 3 depicts a system to record postural variability of a subject according to one embodiment.

[0015] FIG. 4 depicts a method of using a single device to optimize applied stimulation according to one embodiment.

[0016] FIG. 5A depicts mediolateral position over time for one trial of an individual.

[0017] FIG. 5B depicts anterior-posterior position over time for the one trial of the individual of FIG. 5 A.

[0018] FIG. 5C depicts planar Euclidean displacement over time for the one trial of the individual of FIG. 5 A.

[0019] FIG. 6 depicts example fluctuation functions for each of postural-fluctuation time series from thirty trials plotted on logarithmically scaled axes.

[0020] The invention will be better understood and aspects of the inventions other than those set forth above will become apparent when consideration is given to the following detailed description thereof.

DETAILED DESCRIPTION OF PREFERRED EMBODIMENTS

[0021] Balance, gait, and other coordinated movements of humans and other mammals rely on the real-time control of muscle contractions in response to volition and changes in the environment. This muscular control is coordinated by the central nervous system (CNS), i.e. the brain and spinal cord, but is reliant on sensory feedback from the extremities. Of primary importance are the mechanical senses that convey knowledge of skin contact with the environment and limb position. [0022] Lack of adequate mechanical sensory feedback is strongly correlated to significant health problems in humans. These include, for example, the tendency of elderly individuals to fall. In fact, falling is a leading cause of injury and subsequent death in the elderly. Thus, there has been a strong need to develop and improve methods and systems for reducing postural variability.

[0023] A collection of specialized sensory cells, called mechanoreceptors, is responsible for providing the flow of sensory information from the extremities. Mechanoreceptors transduce mechanical stimuli from bodily movements and interactions with the environment into electrical signals that can be transmitted and interpreted by the nervous system. Mechanoreceptors of various types, and found in various anatomic structures, have been identified by researchers in this field. For example, Pacinian corpuscles and other related sensory neurons— found in the skin and deeper subcutaneous layers— are sensitive to touch, stretch, pressure, and vibration ("tactile sense"). Other types of mechanoreceptors (e.g., Golgi tendon organs and muscle spindles) are found in tendons, ligaments, muscles, and tissues within joints. They convey information about muscle force, muscle stretch, and joint angles ("joint sense" or "proprioception").

[0024] Mechanoreceptors are threshold-based units. That is, the cell will not activate and begin signaling unless the magnitude of an environmental stimulus reached a threshold level. A stimulus that is below this level is called "subthreshold" and a stimulus above this level is called "suprathreshold."

[0025] Many health conditions and diseases (e.g., aging, diabetes, stroke, neuropathies, trauma and injury, etc.) can negatively impact either the sensitivity of the mechanoreceptors themselves, the transmission of nerve impulses (action potentials on axons), or the interpretation of nerve impulses centrally at the level of the spine or brain. Lost mechanoreceptor sensitivity is essentially equivalent to a rise in the threshold level.

[0026] It is possible to improve the sensitivity of mechanoreceptors using particular forms of mechanical and/or electrical stimulation applied to the tissue in which the mechanoreceptors are found. This is typically referred to as "stochastic resonance." The applied stimulation should be strong enough to not be damped out, but weak enough to not cause discomfort to the tissue or to otherwise risk harm to the subject. The stimulation is typically applied as temporally uncorrected patterns or "white noise." In one nonlimiting example, stochastic-resonance insoles can be used to lower the magnitude of environmental stimulation required to trigger the mechanoreceptors of the foot by providing subthreshold vibratory mechanical stimulation to the plantar surface of the foot. Several devices to apply stochastic resonance stimulation to the foot are disclosed in Harry et al., U.S. Publication No. 2004/0173220 (filed Mar. 8, 2004), which is hereby incorporated by reference in its entirety. The insoles reduce postural variability by providing white-noise stimulation that is strong enough to not be damped out, but weak enough to not interfere with stability.

[0027] Surprisingly, the effects of stochastic resonance are also affected by endogenous fractal fluctuations. That is, small changes within a person's body, tissue, or even cells affect the efficacy of stochastic resonance stimulation and lead to strong individual differences in non- stimulated measures. These differences can be measured using, for example, non- stimulated posture, inter-stride intervals, regular time-interval estimation, interbreath intervals, and/or inter-heartbeat intervals. Thus, the stabilizing effects of subthreshold vibratory stimulation on, for example, posture depend on baseline variability due to endogenous fractal fluctuations (e.g., the amount of postural variability a person exhibits without stimulation). It is contemplated that measurements of one physiological system can be used to improve efficacy of stimulation on other physiological systems. For example, measurements of interbreath intervals may be used to improve the efficacy of stochastic-resonance stimulation to improve posture.

[0028] Inter-stride intervals can be measured using a number of methods and devices including a pedometer worn on the body, accelerometers and/or pressure pads incorporated into footwear, etc. Regular time-interval estimation can be measured using a number of methods and devices including having the user estimate one-second intervals by repeatedly tapping their finger or foot on a pressure pad, repeatedly clicking a computer mouse, repeatedly pressing a computer key, etc. Interbreath intervals can be measured using a number of methods and devices including a respiratory mask or respiratory inductance plethysmography. Inter-heartbeat intervals can be measured using a number of methods and devices including an electrocardiogram or heart-rate monitoring device.

[0029] Postural variability can be measured using a number of factors. For example, the mean radius (R-Mean), maximum radius (R-Max), elliptical area (A_Eui_pse), mediolateral range (Range_ML), anterior-posterior range (Ranges), mediolateral root-mean-square (RMS_ML), anterior-posterior root-mean-square (RMSA_P), mediolateral path length (Patli_ML), anterior- posterior path length (Patli_Ap), sway speed, and swept area (As_wept) each describe a particular aspect of variability. It is contemplated that other measures may be used. [0030] A radius refers to the absolute distance from the average position during standing. Rt fean is the average distance from the average position during standing. R_Max refers to the maximum absolute distance from the average position during standing. Asiii_pse is the area of an analytically-derived ellipse using the major and minor axes of postural sway. Range_ML is the difference between the leftmost and rightmost excursions of position along the coronal plane. Range_Ap is the difference between forward-most and rear-most excursion of position along the sagittal plane. RMS_ML is the standard deviation of position along the coronal plane. RMSA_P is the standard deviation of position along the sagittal plane.

[0031] Path_ML, Patli_Ap, sway speed, and A_Swept summarize the movement of posture from sample to sample of the recording device. Path_ML is the sum of total absolute Euclidean displacement of postural position along the mediolateral axis. Patli_Ap is the sum of total absolute Euclidean displacement of postural position along the anterior-posterior axis. Sway speed is the average planar Euclidean displacement between each consecutive pair of samples. This can be calculated using, for example, the average square root of the sum of squared displacements along the mediolateral and anterior-posterior axes. As_WePt is a sum of the areas between each consecutive pair of radii. The areas are calculated, for example, as half the cross product of each consecutive pair of radii.

[0032] For the purposes of the present description, the unit of A_Swept and A_E11i_pse is squared millimeters, the unit of sway speed is millimeters per second, and the unit for all other measures is millimeters unless specifically stated. It is contemplated that other units may be used. It is further contemplated that two different units may be used for measurements (e.g. As_wept measured in squared millimeters and A_Eui_pse measured in squared inches) so long as these differences are accounted for.

[0033] These postural-variability measures correlate to the likelihood a person falling while walking or during quite standing. As each of the foregoing postural-variability measures increases, so does the likelihood of a fall. Thus, a subthreshold vibratory stimulation should reduce at least one of the postural variability factors and should also lead to a decrease in at least one of the postural variability measures.

[0034] Other exogenous fluctuations also lead to changes in externally measurable variability. Exogenous fluctuations include all sources external to a person's body such as changes in rigidity of a surface in contact with the body, changes in temperature, etc. [0035] While the division between endogenous and exogenous fluctuations may not be straightforward, it is important to recognize that the stochastic resonance stimulation is not being applied to a dormant, passive system. Rather, endogenous fractal fluctuations result in externally measurable fluctuations that can be temporally correlated. The effects of stochastic resonance can be optimized to suit these temporal correlations. In one nonlimiting example, temporally correlated postural sway can be treated by applying a temporally anti- correlated stochastic resonance stimulation pattern. A crucial point is that endogenous fractal fluctuations are meaningfully related to the effects of externally measurable fluctuations.

[0036] The baseline-dependence of the stochastic-resonance effect on posture is one expression of the influence of endogenous fractal fluctuations. It is contemplated that the relationship may be deeper and more subtle. For example, postural sway is temporally correlated. Changes in temporal correlations of fluctuations in perceptual-motor behavior help to predict individual differences in perceptual and cognitive responses to environmental stimuli. Indeed, it has been suggested that stochastic-resonance control systems in posture may operate on the interaction of temporal correlation in stochastic stimulus with temporal correlation in stimulated system. That is, individual differences in how the human body responds to exogenous fluctuations can be predicted from the temporal correlations of endogenous fractal fluctuations. These considerations may have strong implications for bolstering the clinical value of stochastic resonance. That is, tuning the exogenous fluctuations from, for example, the insole to the temporal correlations of endogenous fractal fluctuations such as postural fluctuations is one way to tailor stochastic-resonance technology to an individual. The superposition or summing together of two or more fluctuations with different degrees of temporal correlations will exhibit a compromise among all superposed fluctuations. The temporal correlation of the superposition is often an amplitude weighted average of temporal correlations.

[0037] Similarly, fitting a temporally correlated fluctuating postural system with a temporally uncorrected foundation (e.g., the vibrating insole) effects a similar compromise. One aspect of the efficacy of the vibrating insole for minimizing postural variability is the de- correlation of postural fluctuations through the superposition of temporally uncorrected noise from the insole. Not only do temporal correlations in postural fluctuations alter the effects of vibrating insoles on postural variability, but the uncorrected noise from vibrating insoles also serves to diminish temporal correlations in postural fluctuations.

[0038] Referring now to FIG. 1, a system 100 for applying neurological stimulation having a temporal variation is shown. The system 100 includes a support platform 102, a bias- signal generator 104, a controller 106, and a power source 108. The bias-signal generator 104 is operatively connected to the support platform 102. The controller 106 controls the bias-signal generator 104. The power source 108 provides electrical energy for the controller 106, the bias-signal generator 104, and the support platform 102. The system 100 can be wearable by an individual and/or incorporated into a single device.

[0039] The power source 108 can be any device capable of delivering power to the system. This can include portable or stationary power sources. Portable power sources include, for example, batteries or fuel cells. The batteries may be single use or rechargeable and include alkaline, nickel cadmium, lithium-ion, polymer, gel, nickel-metal-hydride, etc. Additionally, power recovery technologies may be incorporated such as photoelectric cells or materials and components that generate usable electrical power from excess energy expended during movement of the individual. Stationary power sources include power from an electrical grid or other power source.

[0040] The support platform 102 includes a plurality of bias-signal inputting mechanisms 110, conductors 112, an interface 114, and a sensor 116. The bias-signal inputting mechanisms 110 apply a mechanical and/or electrical stimulation to mechanoreceptors. One nonlimiting example of mechanical stimulation uses a single actuator or combination of actuators such as electromagnetic, electromechanical, solid-state actuators (e.g. Nitiol, piezoelectric), hydraulic, pneumatic, ferrofluid, electroactive polymer, etc. One nonlimiting example of electrical stimulation uses disposable, reusable, and/or stick-slip electrodes. The bias signal inputting mechanisms 110 can be disposed within the support platform 102 and/or on a surface of the support platform 102. The conductors 112 carry signals between both the bias signal inputting mechanisms 110 and the interface 114 and the sensor 116 and the interface 114.

[0041] The interface 114 can be used as a power interface to deliver power to the support platform 102 and/or can be used as a communication interface for coupling a controller with an external device for remote external control, programming, or other purposes. The communication interface can be wired, wireless, or optical. The power interface may be wired or wireless.

[0042] The sensor 116 may be used to collect physiological data from an individual. The sensor 116 may measure force, position of force, acceleration, pressure, etc., or any combination thereof. The sensor 116 transmits information to the controller 106 using the inter- face 114. It is contemplated that some embodiments can include a plurality of sensors while others will not include sensors.

[0043] The bias signal generator 104 generates a driving signal that is supplied to the bias-signal inputting mechanisms 110 using the interface 114. The bias signal inputting mechanisms 110 can be driven individually, in groups, or as a single unit. The driving signal can have a temporal variation.

[0044] Additionally or alternatively, the sensor 116 can be configured to register footfalls during gait over an arbitrary span of time. This span of time could be every day, every week, a set period of time after the user begins using the support platform 102, an interval beginning at a predetermined time of day, etc. The span of time can be selected to exclude long periods with no steps. This data can be collected and analyzed to produce a time series of intervals between steps or strides to determine the degree of temporal correlation. This can be accomplished using a single device or multiple devices. Multiple devices can communicate data using standard methods such as wired or wireless communications. In one nonlim- iting example using multiple devices, the sensor 1 16 would record a series of footfalls while the user wears the support platform 102. When the user charges the system 100 in a charging station, the system 100 transfers the recorded series of footfalls to the charging station. The charging station includes a processor programmed to analyze the degree of temporal correlation in the recorded series and communicate an adjusted stimulation profile to be applied by the system 100.

[0045] FIG. 2 depicts a method of determining a threshold sensory performance for an individual according to one embodiment. A stimulus is applied to the individual at step 202. After applying the stimulus, the individual's response to the applied stimulus is measured at step 204. A determination 206 is made whether the applied stimulus is a sensory threshold. If the applied stimulus is not the sensory threshold, step 208 adjusts the stimulus and the process is repeated until the sensory threshold is determined. If the applied stimulus is the sensory threshold, threshold parameters are set at step 210. Optionally, step 212 communicates these parameters to a wearable device.

[0046] FIG. 3 depicts a system to record postural variability of a subject according to one embodiment. The system includes a main controller 302, a stimulation controller 304, a stimulating device 306, and postural sensors 308. The main controller 302 is operatively connected to the stimulation controller 304 and determines the type and level of stimulation to be applied to the subject during an assigned task. The stimulation controller 304 may apply no stimulation, subthreshold stimulation, or suprathreshold stimulation during the task. The stimulation controller 304 applies the stimulation to the subject using the stimulation device 306. The stimulation device 306 includes at least one bias-signal inputting mechanism to deliver stimulation to mechanoreceptors of the foot.

[0047] The postural sensors 308 are used to record data related to the postural variability of the subject while the subject performs an assigned task such as, for example, quiet standing. Nonlimiting examples of postural sensors 308 include, pressure sensors, force sensors, accelerometers, positioning sensors, etc. Positioning sensors that may be used include pressure sensors in contact with the plantar surface of the individual's feet or, for example, a motion capture system that records the location and/or position of a marker placed on the individual's body. The data recorded by the postural sensors 308 is relayed to the main controller 302. As before, the data can be used to determine stimulation parameters and/or a stimulation profile that may be delivered to a wearable stimulation system 310.

[0048] Position and velocity information in postural sway can be exhausted using data from anterior-posterior position, mediolateral position, and planar Euclidean displacement. The planar Euclidean displacement time series is not a typical measure of postural sway. The planar Euclidian displacement is essentially the first-order difference of center-of-pressure path length. The planar Euclidean displacement time series is analogous to the absolute Euclidean displacement time series which provides information regarding the role of fluctuations in exploratory behaviors. As will be described in further detail below, the planar Euclidian displacement carries independent information (e.g., related to velocity) that cannot be found in either the anterior-posterior and/or the mediolateral position time series alone.

[0049] The anterior-posterior position, mediolateral position, and planar Euclidean displacement time series are analyzed using detrended fluctuation analysis to determine scaling exponents H for each series. It is contemplated that other methods such as dispersion analysis, stabilogram diffusion analysis, power-spectrum analysis, and rescaled-range analysis, etc. may be used to analyze the time series. Detrended fluctuation analysis is an adaptation of a random-walk analysis that examines the growth of root-mean-square fluctuations over the course of a time series while also controlling for nonstationarities due to drift as described in Peng, C. K. et al, Mosaic Organization of DNA Molecules, Phys. Rev. E, 49, 1685-1689 (1994), which is incorporated by reference in its entirety. Although detrended fluctuation analysis is conceptually related to stabilogram diffusion analysis such as that described in Duarte & Zatsiorsky, 2000, 2001 , there are algorithmic differences that informed selection of detrended fluctuation analysis. One difference is that detrended fluctuation analysis conservatively removes drift such as artifactual trends and/or spurious trends before assessing fluctuations, discussed in, for example, Delignieres, D., et al, Transition from Persistent to Anti- persistent Correlations in Postural Sway Indicates Velocity-based Control. 7 PLoS Computational Biology el 001089 (201 1), which is hereby incorporated by reference.

[0050] The intent of detrended fluctuation analysis is to estimate a scaling exponent H to index temporal correlations. The analysis begins with the integration of a time series x(t) into a random- walk trajectory y(t), as follows:

N

y(t) =∑x(i) (1) where x( ) is the mean of x(t). Detrended fluctuation analysis calculates the root-mean- square after removing local trends. Linear regressions of y„(t) detrend non-overlapping n- length bins of y(t). Fluctuation F(n) is calculated as average root-mean-square error of these regressions for each n as follows:

typically for n < N 1 4. However, F(n) can be unstable for larger n because there are relatively fewer bins for relatively larger bin sizes. If desired and/or necessary, detrended fluctuation analysis can be run conservatively by limiting F(n) to n < N I 10. Although the error due to the scaling region may be negligible for individual estimates of the scaling exponent H on a given trial, the present strategy of generating multiple estimates of the scaling exponent H using multiple fluctuation time series for the same trial would risk compounding this error. Thus, for the present disclosure, estimates of scaling exponents H are limited to the more conservative scaling region, n < N I 10. F(n) increases over the scaling region as:

H

Fin) n (3)

Logarithmic scaling of Equation 3 yields:

Thus, the slope of (n) in a double-log plot is taken to estimate H. [0051] Temporally correlated empirical time series generally have scaling exponents H greater than 0.5 whereas temporally uncorrected empirical time series generally have scaling exponents H equal to 0.5. However, it is possible for distributional anomalies to produce scaling exponents spuriously greater than 0.5 in the absence of temporal correlations. In order to test that a sufficiently large scaling exponent H reflects temporal correlations, detrend- ed fluctuation analysis is run on a shuffled time series. The shuffled time series is includes the same values as the original time series, but the values are arranged in a random order. According to theory, the shuffled time series should have a scaling exponent H equal to 0.5. In practice, sampling error and/or departures from pure normality may lead to scaling exponents that only approximate a scaling exponent of 0.5. Even if the shuffled scaling exponent H does not equal 0.50, a temporally correlated original series scaling exponent H should exceed the shuffled scaling exponent H.

[0052] Growth curve modeling may be used to analyze postural-variability measures and to test whether the temporal correlation of endogenous postural fluctuations affect the negative effects of subthreshold stimulation on postural-variability measures. The endogenous postural fluctuations used for the analysis may be represented by, for example, the estimated scaling exponents H. Growth curve modeling is a longitudinal, maximum-likelihood multiple-regression technique designed to test the effect of time-varying predictors and is well- suited for testing the time-varying effects of endogenous physiological fluctuations on a biological organism's response to stimulation. It is contemplated that other regression methods such as those using ordinary-least-squares estimation may be used.

[0053] Growth curve models decompose a dependent measure in terms of a weighted sum of linearly separable predictors, and return estimates of coefficients for each predictor. An important difference between a growth curve model and an ordinary least squares regression technique (e.g., RM ANOVA) is the assumptions about the distribution of error over time. Ordinary least squares estimation assumes equal variance over time and across participants. However, growth curve modeling uses maximum likelihood estimation to fit random effects for individual differences across participants and over time. The effect of added predictors in ordinary least squares estimation is evaluated in terms of a change in proportion of explained variance (e.g., i?-squared). In contrast, the maximum likelihood estimation for a continuous dependent measure allows no absolute goodness-of-fit statistic and, therefore, no reliable description of proportion of explained variance. Instead, nested models can be evaluated based on the reduction of a -2 log likelihood deviance statistic. Improvement in model fit following the addition of m new parameters is evaluated in terms of -2 log likelihood deviance, where change in -2 log likelihood is tested as a chi-square statistic with m degrees of freedom.

[0054] A power spectrum relates power P, frequency and power-law exponent β as follows:

Ρ(/) ~ Γ^β (5)

White-noise or uncorrected noise has a power-law exponent β of zero, correlated noise has a power-law exponent β greater than zero, and anti-correlated noise has a power-law exponent β less than zero. The scaling exponent H is related to the power-spectral power-law exponent β as:

H (6)

Rearranging Equation 6 yields: β = 2Η - 1 (7)

[0055] The marginal difference βά is calculated to counteract deviation from the average power-spectrum power-law exponent. This is done by subtracting a mean estimated power- spectral power-law exponent k from the power-spectral power-law exponent β. That is:

(8)

Thus, the marginal difference βά highlights the excursion of a given participant from typical power-law structure estimated from measurements of a given group of individuals and/or from repeated measurements of the given participant. For example, a pilot study, which will be described in greater detail below, found the mean estimated power-spectral power-law exponent k for young adults is 0.66 and the mean estimated power-spectral power-law exponent k for elderly adults is 0.76. This indicated a steeper power law for the planar Euclidean displacement of elderly individuals than young individuals. It is contemplated that other constants may be determined or that a more robust term may be added.

[0056] A noise waveform can be calculated and a drive signal programmed to deliver a power spectrum P(f) that increases as a function of frequency For example,

P(f) ~ " (9) The maximum power delivered by the power spectrum is set to the maximum power of an original white-noise power spectrum. Typically, the maximum power of the original white- noise power spectrum is ninety percent (90%) of the individual's sensory threshold. It is contemplated that other values may be used. When the marginal difference βά is greater than zero, the maximum power delivered occurs at the highest available frequency f_meiX (e.g., P(fmnx))- Thus, the calculated noise waveform and drive signal should de-correlate endogenous fractal fluctuations when the marginal difference β^// is greater than zero. When the marginal difference ?<¾ is less than zero, the maximum power delivered occurs at the lowest available

(e.g., P(fwn))- Thus, the calculated noise waveform and drive signal should strengthen correlations of endogenous fractal fluctuations when the marginal difference β diff is less than zero. In both cases, the goal of the recalculated noise signal is to bring the endogenous fluctations' power-spectral power-law exponent closer to the mean estimated power-spectral power-law exponent k. Additionally, the ratio between the maximum power delivered and the minimum power delivered will depend on the range of frequencies available.

[0057] In an embodiment, a single wearable device can measure temporal correlations in planar Euclidean displacement, estimate the marginal difference βα -, and recalculate the drive signal as above. The wearable device includes a bias-signal inputting mechanism, a bias-signal generator, a sensor, a controller, and a power source. The power source supplies power to the device and is preferably portable. The controller is operatively connected to the power source, the sensor, and the bias-signal generator. The bias-signal generator is operatively connected to the bias-signal inputting mechanism.

[0058] One method of using the single device is illustrated in FIG. 4. An individual wears the single wearable device at step 402. While wearing the device, the individual performs a specified task at step 404. This task may be, for example quiet standing for a specified period of time such as thirty seconds. The specified task can be performed with no stimulation, subthreshold stimulation, or suprathreshold stimulation. Step 406 acquires data regarding postural fluctuations using the sensors during the specified test. This data may include the pressure exerted on certain points of the plantar surface over the duration of the specified task. The acquired data is processed by the microprocessor at step 408. The microprocessor may work up the data using algorithms discussed above such as detrended fluctuation analysis to estimate the marginal difference and recalculate the drive signal. [0059] After the method of FIG. 4 has recalculated the drive signal, a single device may be used to stimulate the mechanical receptors of the individual using the recalculated drive signal. The method of FIG. 4 can be performed at various intervals such as daily, weekly, monthly, yearly, as needed, or as desired. It is contemplated that the sensors may constantly collect information during normal use of the device. This would allow the controller to contemporaneously change the applied stimulation to optimize performance.

[0060] The actuators discussed thus far have been active actuators that require an electrical power source and driving signal to provide a stimulating vibration to a mechanoreceptor site. However, the invention is not limited to the use of active devices. Passive vibrational actuators may also be used. Passive mechanical actuators are constructed from materials that generate mechanical vibrations as they are compressed by body weight during locomotion, etc. Such mechanisms incorporate a bias structure that returns the actuator to its original position when the load is removed. As compression or decompression takes place, the actuator emits a vibration. That is, during striding, the passive actuator structure is repeatedly compressed by the application of body weight, and returned to its original position. Consequently, useful mechanical vibrations are generated.

[0061] The efficacy of the vibrating insoles for stabilizing posture lies in a compromise between the temporal correlations of intervention and the physiological fluctuations. Unlike many other clinical interventions, vibrating insoles have been explicitly designed to generate fluctuations with a specific degree of temporal correlation. Temporal correlations may serve as a common currency to understand the relationship between vibrating insole and postural system because physiological fluctuations are temporally correlated. Temporal correlations in biological systems can vary widely across an individual's lifespan. Biological systems appear to fare best when fluctuations are temporally correlated but not excessively so. It may be that fractal ("1/f") fluctuations (e.g., "pink noise") may be beneficially used because they reflect a power-law balance between more random and uncorrected fluctuations (e.g., "white noise") and more determined and correlated fluctuations (e.g., "Brownian noise").

[0062] The concepts of the present disclosure appear to reflect a similar type of balance. First, it appears that increases in postural-variability measures are associated with increases in temporal correlation of planar Euclidean displacements in posture fluctuations. Second, it appears that an interaction between temporal correlations of planar Euclidean displacements and temporal correlations of mediolateral position moderates the effect of white-noise insole vibrations. Third, it appears that elderly individuals exhibit an increase in temporal correla- tions. Finally, it appears that stimulation by vibrating insoles diminishes the same temporal correlations.

[0063] It is contemplated that one type of stimulation or noise may not be optimal for all patients. Optimization of stimulation may involve different degrees of temporal correlation for different systems on different time scales. Different physiological systems may benefit more suited for white noise, pink noise, Brownian noise, or even a combination of these.

[0064] The time scale of the intervention may also have an effect on the type and amount of noise applied. For example, anterior-posterior postural sway in the healthy elderly can sometimes exhibit weaker temporal correlations than in healthy adults over more prolonged standing periods.

[0065] While the foregoing specific embodiments of the present invention as set forth discuss applicability of the present concepts to improve posture and gait, it is contemplated that other physiological systems may be improved using the concepts described herein without departing from the main theme of the invention. By way of nonlimiting example, these physiological systems can include cardiovascular or respiratory systems.

EXAMPLES

[0066] Postural variability measures were analyzed using the above disclosed methods. The analysis showed that postural-variability measures increased with increases in the temporal correlation of planar Euclidean displacements. The analysis further showed a reduction of postural variability due to insole vibrations that was moderated by the interaction of temporal correlation of planar Euclidean displacements with temporal correlation of mediolateral position. Additionally, the analysis found that elderly planar Euclidean displacement exhibited power-spectra decaying according to negative power-law functions of frequency. The power-law exponents for elderly subjects were found to be about 0.1 lower than those of younger patients. This indicated a steeper power law. The average power-spectrum power- law exponent for young adults was 0.66. The average power-spectrum power-law exponent for elderly adults was 0.76. These numbers were obtained using a Wiener-Kinchin transformation of Hurst exponents.

[0067] FIGS. 5A-5C show time series measurements for three postural-variability measures during an example 30-second trial that were analyzed using the above methods. FIG. 5A depicts mediolateral position over time for one trial of an individual. The mediolateral position begins at about 250 millimeters at the start of the time series and steadily in- creases to about 270 millimeters at about three seconds. The position remains fairly level at about 270 millimeters until about fifteen seconds. The position then begins to oscillate between about 260 millimeters and 280 millimeters between fifteen seconds and the end of the time series.

[0068] FIG. 5B depicts anterior-posterior position over time for the one trial of the individual of FIG. 5 A. The anterior-posterior position begins at 140 millimeters and fluctuates between about 135 millimeters and about 150 millimeters until six seconds. The positions slowly descend from about 150 millimeters to about 145 millimeters between six seconds and fifteen seconds. The anterior posterior position fluctuated between about 150 millimeters and about 132 millimeters between about fifteen seconds and about twenty-five seconds. After twenty-five seconds the anterior posterior position began to slowly rise from about 140 millimeters to about 160 millimeters.

[0069] FIG. 5C depicts planar Euclidean displacement over time for the one trial of the individual of FIG. 5A. The planar Euclidean displacements fluctuate between about 0.0 and about 0.6 while remaining relatively consistent across the time series. Some features can be seen between about ten seconds and about fifteen seconds and between about fifteen seconds and about twenty seconds. The fluctuations slow between about ten seconds and about fifteen seconds. The fluctuations then grow between about fifteen seconds and about twenty seconds. As shown, the planar Euclidean displacement fluctuations at the end of the time series were generally greater than the fluctuations of the beginning of the time series

[0070] Multiple 30-second trials on multiple subjects were analyzed using fine-grain postural fluctuations for the measurements of FIGS. 5A-5C. Each time series was recorded by placing a near-infrared-refiective marker on the right shoulder of a participant and recording the marker position with a VICON motion-capture system. The postural data was sampled at 60 Hz. Further details of data collection may be found in Priplata, A. A. et al., Vibrating Insoles and Balance Control in Elderly People, 362 The Lancet 1123-24 (2003), which is herein incorporated by reference for its method of data collection and its obtained data. The illustrated time series were derived from the recorded VICON data.

[0071] The Mediolateral Sway time series measured the excursion from average position along the sagittal plane for the example trial. The Anterior-Posterior Sway time series measured the excursion from the average position along the coronal plane for the example trial. The Planar Euclidean Displacement time series is the square root of the sum of squared ex- cursions for the example trial. It is contemplated that different, additional, and/or fewer measures of postural variability may be used.

Testing Time-dependence

[0072] As discussed previously, detrended fluctuation analysis was used to provide an estimate of scaling exponent H for each of the three postural-fluctuation time series on each trial. Estimates of the scaling exponent H for an anterior-posterior position time series is referred to as HAp. Estimates of scaling exponent H for a mediolateral position time series is referred to as HML- Estimates of scaling exponent H for planar Euclidean displacement time series is referred to as HPED-

[0073] Detrended fluctuation analysis was performed on each of 20 trials from 15 young participants and 10 trials from 12 elderly participants. This yielded a total of 1260 scaling exponents H (i.e., 20* 15*3 + 10* 12*3). Each scaling exponent HAp, HML, and HPED was estimated for both the original postural-fluctuation time series and for shuffled copies. Table 1 shows the mean and the standard error for each scaling exponent estimated using the trial data.

Table 1 - Scaling exponents for original and shuffled time series

HAP HML HPED

M SE M SE M SE

Original 1.81 0.01 1.78 0.01 0.85 0.01

Shuffled 0.50 0.01 0.50 0.01 0.53 0.01

[0074] As discussed above, the scaling exponents H were estimated using detrended fluctuation analysis only for the scaling region within a conservative bound where bin sizes n < N I 10, and scaling exponents for the shuffled copies were also calculated in order to determine that the scaling exponents were due to distributional anomalies. As shown in Table 1 , the estimated scaling exponents H of each time series exceeded the estimated scaling exponent for the corresponding shuffled time series where M is the mean of the scaling exponent H and SE is the standard error. Example fluctuation functions for each of the postural- fluctuation time series have been plotted on logarithmically scaled axes in FIG. 6.

Growth Curve Modeling

[0075] For the purposes of this disclosure, growth curve models are described in terms of the highest-order interactions. This description is made with the understanding that all lower- order interactions and main effects are necessarily included concurrently to permit their standard interpretation. In order to test whether endogenous postural fluctuations moderated the negative effect of subthreshold vibratory fluctuations on postural variability, each postural-variability measure was modeled with the same set of predictors. A predictor representing insole stimulation Stim is coded as 0 for trials with no subthreshold vibratory stimulation and 1 for trials with subthreshold vibratory stimulation. Similarly, a predictor representing age group Age was coded as 0 for young adult participants and 1 for elderly participants. The predictors representing endogenous postural fluctuations were the trial-by-trial values of HAP, HML, and HPED as estimated by detrended fluctuation analysis. Specifically, the highest-order term in a first model was Stim* Age* HAP*HML*HPED, all constituent lower-order interactions and main effects thereof, and a main effect of Trial. The interaction of stimulation, age group, and temporal correlation of postural fluctuations on postural variability was modeled by the interaction Stim* Age* HA_P *H_ML*H?_EO and all constituent terms. The main effect of Trial controlled for effects of time spent in the task (e.g., fatigue). Additional modeling suggested that interactions of Trial with the Stim *Age *HAP *HML *HPED term did not significantly improve model fit. Interactions of Trial with other predictors may support improvements in model fit in tasks with closer to one-hundred trials, but tasks with approximately twenty trials needed to only incorporate Trial in terms of a standalone covariate for time spent in the task.

[0076] One more statistical step was used to address collinearity of effects in the predictors and thus provide interpretable growth curve models. As discussed above, each postural- fluctuation time series carried substantial independent information about postural sway. The simple pairwise correlations among the scaling exponents HAp, HML, and HPED supported this because the correlations were relatively weak as can be seen in Table 2.

Table 2 - Correlation matrix for estimates of scaling exponents

HAP HML HPED

1.00 0.2 0.30

1.00 0.18

As shown, the maximum correlation between any two different estimated scaling exponents was 0.30. However, the introduction of these terms as main effects and as part of variously higher-order interactions into the same model produced strong collinearities. That is, even when main effects (e.g., scaling exponents H) are not themselves strongly correlated, predictors representing higher-order interactions may be correlated. As with any other multiple regression technique, growth curve modeling estimates may be misleading if predictors are cor- related. Efforts were taken to orthogonalize those main effects participating in higher-order interactions.

[0077] The first model,

+ Trial, tested the effects of interactions among subthreshold vibratory stimulation Stim, age group Age, and temporal correlations in postural fluctuation as well as an effect of the trial Trial on the magnitude of postural-variability measures. This model included thirty-one total predictors, but these predictors exhibited a preponderance of very high correlation with one another (i.e., r > 0.9).

Table 3 - All thirty-one predictors tested

• Stim • Stim*HAp • HAP*HPED • Stim*H_ML*HpED

• Age • Stim*H_ML • HML*HPED • Age * HAP * HML

• HAP • Stim*HpED • Stim*Age*HAP • Age*HAp*HpED

• H_ML • Age* HAP • Stim*Age*H_ML • Age*H_ML*HpED

• HpED • Age*H_ML • Stim*Age*HpED • HAP*HML*HPED

• Trial • Age*Hp_ED • Stim*HAP*H_ML • Stim*Age*HAP*H_ML

• Stim* Age • HAP*HML • Stim*HAP*HpED • Stim*Age*HAP*HpED

• Stim*Age*H_ML *HPED · Age*HAP*H_ML*HpED · Stim*Age*HAP*H_ML*HpED

[0078] To overcome collinearity among the predictors, the three scaling exponents HAp, H_ML, and HpED were replaced with the corresponding three principal components PCi, PC₂, and PC₃. This step dramatically attenuated the correlations among fixed effects, with correlations remaining relatively weak (i.e., approximately 90% of correlations with absolute-value of r < 0.25 and no correlations with absolute-value of r > 0.9). Table 3 shows the loadings for the principal components.

Table 4 - Loadings of scaling exponents HAp, HML, HPED, on PC_1; PC₂, and PC3

PCi PCj PCs

0.28 0.00 -0.96

0.35 0.92 0.20

0.89 -0.39 0.23

These loadings indicated that Η_ΡΕΌ, H_ML, and HAP contributed most to PCi, PC₂, and PC3 respectively. Reinforcing this point, there were strong correlations between HpED and PCi, (r(418) = 0.94, /? < .0001), H_ML and PC₂ (r(418) = .88, /? < .0001), and HAP and PC₃ (r(418) = -0.87, p < .0001). [0079] The second model was preferred to the first model, including

all lower-order interactions and main effects thereof, and the main effect of Trial.

Testing Temporal Correlations of Postural Fluctuations.

[0080] Each of the eleven postural-variability measures was analyzed using growth curve models that contained the predictors for the second model outlined above. It was found that the coefficients for Stim, Age, PCi, and PCi*PC₂*S¾^'m were either most theoretically salient (e.g., Stim and Age) or exhibited significant effects most consistently across postural- variability measures (e.g., Stim, PCi, PCi*PC₂*S¾^'m). Significantly large coefficients for each of these predictors indicated significant effects of subthreshold vibratory stimulation Stim, age group Age, the first principal component, and the interaction of the first two principal components of the scaling exponents and subthreshold vibratory stimulation. These calculated coefficients are illustrated in Table 4 for each of the eleven postural variability measures.

Table 5 - Determined coefficients for eleven postural variability measures

Stim Age PCi PCi*PC₂*Stim

R-Mean -0.63 0.06 9.19^"* - 149.8Γ

R-Max - 1.14 0.68 17.66 -351.97^*

****

-^Elli se -52.5 72.7 981.10 - 1 1348.4^*

Range_ML -0.19 1.38 20.32^** - 134.85

RangeAp -2.61 1.35 29.56 -661.65

RMSML -0.19 0.06 6.423^*** -39.20

RMSAP -0.67 0.15 7.45 - 189.42

Path_ML -2.05 38.28 90.48 - 1668.97^*

PatliAp -3.02 42.42 108.70^*** - 1487.72^*

Sway speed -0.14 2.14 5.27 -79.50^*

-A-Swept -70.87^* 182.67 1 139.54^*** - 18512.85^*

Superscripts *, **, ***, and **** denote ps < .05, .01, .001, and .0001, respectively.

[0081] As shown in Table 4, each of the eleven postural variability measures had a consistent direction across each of the predictors. That is, stimulation Stim and PCi*PC₂*S¾^'m had negative effects for each postural-variability measure, whereas age group Age and PCi had positive effects. These consistencies confirmed that subthreshold vibratory stimulation reduces postural variability. The consistencies of direction also supported subthreshold vibratory stimulation moderating changes in temporal correlation of postural fluctuations. These postural fluctuations were represented by the three principal components of the three scaling exponents HAp, HML, and HPED- [0082] Specifically, the first principal component PC i generally contributed to postural variability. Further, the interaction of the first two principal components PCi,PC₂ with subthreshold vibratory stimulation Stim generally reduced postural variability.

[0083] The effects supporting previous findings were less consistently significant. The effects of stimulation Stim were significant on only six of eleven postural-variability measures. The postural-variability measures that stimulation Stim affected were R-Mean, Rfviax, A_Eiiipse, RangeAP, RMSAP, and As_wept. The effects of age group Age were significant on only two of eleven postural-variability measures. The postural-variability measures that age group Age affected were Patli_Ap and sway speed.

[0084] Surprisingly, effects implicating the role of temporal correlations in postural fluctuations were more consistently significant. For example, the effects of PCi were significant on all eleven postural-variability measures. The effects of PCi*PC₂*iS¾^'m were significant on nine of eleven postural-variability measures. The postural-variability measures that PCi*PC₂*,S¾^'m affected were R_Mean, Rtviax, A_En_ipse, Range_Ap, RMSAP, Path_ML, PatliAp, sway speed, and A_Swe_Pt.

[0085] The growth curve models demonstrated two points. First, increases in the first principal component PCi predicted significantly greater postural variability and helped predict the magnitude of postural sway. Increased temporal correlations of all aspects of postural sway predicted increased postural sway because all scaling exponents HAP, HML, and HpED had positive loadings on the first principal component PCi. The temporal correlations of planar Euclidean displacement may have had a slightly larger role than those of anterior- posterior or mediolateral position because the dominant loading on the first principal component PCi was HpED- Second, the significant PCi*PC₂*iS¾^'m effect indicated that the interaction of the first two principal components PCi, PC₂ moderated the negative effect of stimulation Stim. When the first two principal components PCi, PC₂ were changed in the same directions, the original effect of stimulation Stim on postural variability is accentuated or made more negative. Conversely, the product is negative when the first two principal components PCi, PC₂ were changed in opposite directions and, therefore, the effect of stimulation of postural variability is diminished or made more positive. This indicates that the temporal correlations of planar Euclidean displacements and of mediolateral sway effect the stimulation Stim because the strongest loadings on the first two principal components PCi, PC₂ were HpED and HML- Effects of Age Group and Stimulation on Temporal Correlations

[0086] Surprisingly, the effects of the first principal component PCi and PCi*PC₂*S¾^'m appeared to be more reliable predictors than either stimulation Stim or age group Age alone. Thus, despite previous findings of differences due to stimulation Stim or age group Age, the differences may have a greater dependence on the temporal correlations in postural fluctuations.

[0087] A new growth curve model was run to test to test for the effects of stimulation Stim and age group Age on the first and second principal components PCi, PC₂ and on PCi*PC₂ to determine whether the traditional predictors of stimulation Stim and age group Age bore any relationship to the more strongly predictive effects of the first principal component PCi and PCi*PC₂*iS¾^'m. It was determined whether the temporal correlations themselves responded to any differences by stimulation Stim or age group Age because of the finding that temporal correlations appeared to influence postural variability directly and to moderate the effect of subthreshold vibratory stimulation. The model tested for differences in PCi, PC₂, and PCi*PC₂ concurrently using a class variable to distinguish significant differences specific to each three dependent variables to control for the relationship among them.

[0088] Some significant effects were a positive effect of age group Age on the first principal component PCi (B = 0.06, SE = 0.02, p < 0.01) and a negative effect of stimulation Stim on PCi (B = -0.014, SE = 0.007, p < 0.05). Thus, Age group Age increased and stimulation Stim decreased values of the first principal component PCi. The correlations between scaling exponents H to principal components PC suggested that age group Age and stimulation Stim should exert these effects on the scaling exponent HpED- Alternate modeling confirms this point by demonstrating that age group Age serves to increase HpED and that stimulation Stim serves to decrease HpED- Thus, age group Age serves to strengthen and stimulation Stim serves to weaken the temporal correlations in the planar Euclidean displacements of posture.

[0089] The subject matter of the present invention can be defined by any of the following paragraphs: A. A system for neurological stimulation comprising: at least one bias-signal inputting mechanism configured to apply a subthreshold stimulation to mechanoreceptors; at least one bias-signal generator coupled to the at least one bias-signal inputting mechanism and configured to provide a driving signal to drive the at least one bias-signal inputting mechanism, the driving signal having a temporal variation; a controller for controlling the at least one bias signal generator and the at least one bias signal inputting mechanism; and a power source providing electrical energy to the controller and the at least one bias signal generator.

B. The system for neurological stimulation of paragraph A wherein the stimulation is a subthreshold stimulation that includes a noise signal.

C. The system for neurological stimulation of paragraph A wherein the temporal variation is according to a predefined temporal profile.

D. The system for neurological stimulation of paragraph A wherein the temporal variation is determined as function of at least one of postural variability, regular time-interval estimation variability, inter-stride interval variability, inter-breath interval variability, and inter-heartbeat interval variability.

E. The system for neurological stimulation of paragraph A wherein the temporal variation is determined as a function of regular time-interval estimation variability, wherein regular time-interval estimation variability is determined by measuring variability of a subject performing at least one of tapping their finger on a pressure pad, tapping their foot on a pressure pad, repeatedly clicking computer mouse and repeatedly pressing a computer key. F. The system for neurological stimulation of paragraph A wherein the temporal variation is determined as a function of postural variability, wherein postural variability is determined by measuring at least one of mean radius (R-Mean), maximum radius (R-Max), elliptical area (A_Eiii_Pse), mediolateral range (RangeML), anterior-posterior range (RangeAp), mediolateral root-mean-square (RMS_ML), anterior-posterior root-mean-square (RMSAP), mediolateral path length (Path_ML), anterior-posterior path length (Patli_Ap), sway speed, and swept area (As_wept).

G. A method for neurological stimulation comprising the acts of: providing at least one bias-signal inputting mechanism configured to apply a subthreshold stimulation to mechanoreceptors, at least one bias-signal generator configured to provide a driving signal to drive the at least one bias-signal inputting mechanism, a controller for controlling the at least one bias signal generator and the at least one bias signal inputting mechanism, and a power source providing electrical energy to the controller and the at least one bias signal generator; and activating the signal generator and supplying a bias signal to stimulate the mechanoreceptors, the bias signal having a temporal variation based on a determined therapeutic need of an individual.

H. The method for neurological stimulation of paragraph G wherein the stimulation is a subthreshold stimulation that includes a noise signal.

I. The method for neurological stimulation of paragraph G wherein the temporal variation is according to a predefined temporal profile.

J. The method for neurological stimulation of paragraph G wherein temporal variation is determined as function of at least one of postural variability, regular time-interval estimation variability, inter-stride interval variability, inter-breath interval variability, and inter-heartbeat interval variability. K. The method for neurological stimulation of paragraph G wherein temporal variation is determined as a function of regular time-interval estimation variability, wherein regular time-interval estimation variability is determined by measuring variability of a subject performing at least one of tapping their finger on a pressure pad, tapping their foot on a pressure pad, repeatedly clicking computer mouse and repeatedly pressing a computer key.

L. The method for neurological stimulation of paragraph G wherein the temporal variation is determined as a function of postural variability, wherein postural variability is determined by measuring at least one of mean radius (R-Mean), maximum radius (R-Max), elliptical area (A_Eui_pse), mediolateral range (Range_ML), anterior-posterior range (Ranges), mediolateral root-mean-square (RMSML), anterior-posterior root-mean-square (RMSAP), mediolateral path length (Path_ML), anterior-posterior path length (Patli_Ap), sway speed, and swept area (A_Swept).

M. A method for developing a stimulation profile comprising the acts of: measuring at least one predetermined factor at a plurality of times, at least two of the times being temporally distinct, the at least one predetermined factor exhibiting temporal fluctuations; formulating a temporal profile using at least two of the temporally-distinct times; and developing the stimulation profile using the temporal profile, the stimulation profile, when applied to mechanoreceptors via at least one bias signal inputting mechanism, improving system stability.

N. The method of paragraph M wherein the stimulation profile includes subthreshold stimulation that includes a noise signal.

O. The method of paragraph M wherein the stimulation profile consists of subthreshold stimulation. P. The method of paragraph M wherein the act of measuring includes using at least one sensor configured to register fluctuations in pressure, the at least one sensor and the at least one bias signal inputting mechanism being disposed in a single device.

Q. The method of paragraph M wherein the predetermined factor includes at least one of postural variability, regular time-interval estimation variability, inter-stride interval variability, inter-breath interval variability, and inter-heartbeat interval variability.

R. The method of paragraph M wherein the predetermined factor includes regular time-interval estimation variability and the regular time-interval variability, wherein regular time-interval estimation variability is determined by measuring variability of a subject performing at least one of tapping their finger on a pressure pad, tapping their foot on a pressure pad, repeatedly clicking computer mouse and repeatedly pressing a computer key.

S. The method of paragraph M wherein the predetermined factor includes postural variability, wherein postural variability is determined by measuring at least one of mean radius (PvMean), maximum radius (R-Max), elliptical area (A_Eiii_Pse), mediolateral range (RangeML), anterior-posterior range (RangeAp), mediolateral root-mean-square (RMS_ML), anterior-posterior root-mean-square (RMSA_P), mediolateral path length (Path_ML), anterior- posterior path length (Patli_Ap), sway speed, and swept area (A_Swept).

[0090] The foregoing specific embodiments of the present invention as set forth in the specification herein are for illustrative purposes only. Various deviations and modifications can be made within the spirit and scope of this invention, without departing from the main theme thereof. It will be appreciated by persons skilled in the art that the present invention is not limited by what has been particularly shown and described herein above.

Claims

What is claimed is: CLAIMS

1. A system for neurological stimulation comprising: at least one bias-signal inputting mechanism configured to apply a subthreshold stimulation to mechanoreceptors; at least one bias-signal generator coupled to the at least one bias-signal inputting mechanism and configured to provide a driving signal to drive the at least one bias-signal inputting mechanism, the driving signal having a temporal variation; a controller for controlling the at least one bias signal generator and the at least one bias signal inputting mechanism; and a power source providing electrical energy to the controller and the at least one bias signal generator.

2. The system for neurological stimulation of claim 1 wherein the stimulation is a subthreshold stimulation.

3. The system for neurological stimulation of claim 1 wherein the temporal variation is a predefined temporal profile.

4. A method for neurological stimulation comprising the acts of: providing at least one bias-signal inputting mechanism configured to apply a subthreshold stimulation to mechanoreceptors, at least one bias-signal generator configured to provide a driving signal to drive the at least one bias-signal inputting mechanism, a controller for controlling the at least one bias signal generator and the at least one bias signal inputting mechanism, and a power source providing electrical energy to the controller and the at least one bias signal generator; and activating the signal generator and supplying a bias signal to stimulate the mechanoreceptors, the bias signal having a temporal dependence based on a determined therapeutic need of an individual.

5. The method for neurological stimulation of claim 4 wherein the stimulation is a subthreshold stimulation.

6. The method for neurological stimulation of claim 4 wherein the temporal variation is a predefined temporal profile.

7. A method for developing a stimulation profile comprising the acts of: measuring at least one predetermined factor at a plurality of times, at least two of the times being temporally distinct, the at least one predetermined factor exhibiting temporal fluctuations; formulating a temporal profile using at least two of the temporally-distinct times; and developing the stimulation profile using the temporal profile, the stimulation profile, when applied to mechanoreceptors via at least one bias signal inputting mechanism, improving system stability.

8. The method of claim 7 wherein the stimulation profile includes subthreshold stimulation.

9. The method of claim 7 wherein the stimulation profile consists of subthreshold stimulation.

10. The method of claim 7 wherein the act of measuring includes using at least one sensor configured to register fluctuations in pressure, the at least one sensor and the at least one bias signal inputting mechanism being disposed in a single device.