System and Method for Prediction of Cognitive Decline

• US 20100094155A1
• Filed: 12/16/2009
• Published: 04/15/2010
• Est. Priority Date: 10/31/2005
• Status: Active Grant

• Current Assignee: New York University
• Original Assignee: New York University
• Inventors: PRICHEP, Leslie S.
• Granted Patent: US 7,986,991 B2
• Time in Patent Office: Days
• US Class Current: 600/544
• CPC Class Codes: A61B 5/246 using evoked responses
• CPC Class Codes: A61B 5/38 Acoustic or auditory stimuli
• CPC Class Codes: A61B 5/4064 Evaluating the brain A61B5/031, A61B5/369, A61B5/14553 take precedence
• CPC Class Codes: A61B 5/4088 Diagnosing of monitoring cognitive diseases, e.g. Alzheimer, prion diseases or dementia
• CPC Class Codes: A61B 5/7264 Classification of physiological signals or data, e.g. using neural networks, statistical classifiers, expert systems or fuzzy systems neural networks per se G06N3/00; expert systems per se G06N5/00
• CPC Class Codes: G16H 50/50 for simulation or modelling of medical disorders

Abstract

A system and method for prediction of cognitive decline, comprises an input receiving input data corresponding to brain activity of an individual and a processor coupled to the input for analyzing the input data to obtain a selected set of features, the processor comparing the selected set of features to at least a portion of entries in a database corresponding to brain activity of a plurality of individuals, wherein entries in the database have been separated into a plurality of categories corresponding to one of a degree of cognitive decline and a propensity for future cognitive decline of individuals relating to the entries, the processor determining, based on the comparison, a category most closely corresponding to the selected set of features.

Claims

1-53. -53. (canceled)

54. A method for prediction of cognitive decline of a patient during a future time period, comprising:

receiving input data corresponding to electrical brain activity of an individual;

analyzing the input data to extract selected features therefrom;

comparing the selected features to at least a portion of entries in a database including data corresponding to brain activity of a plurality of individuals, wherein entries in the database are separated into a plurality of categories corresponding to a degree of cognitive decline of the individuals when the data was collected and a degree of cognitive decline of the individuals during a period of time subsequent to the collection of data; and

determining a category most closely corresponding to the input data.

55. The method according to claim 54, wherein the input data includes one of quantitative electroencephalogram data, quantitative magnetoencephalogram data and quantitative event related potential data.

56. The method according to claim 54, wherein the categories correspond to rankings on a clinical staging scale.

57. The method according to claim 56, wherein the clinical staging scale is a Global Deterioration Scale (GDS).

58. The method according to claim 54, wherein the input data includes ERP data corresponding to time frames immediately after presentation of a series of stimuli presented in one of a plurality of sensory modalities, a plurality of sensory attributes and a plurality of stimulus attributes within a single sensory modality, including common stimuli and at least one of rare and novel stimuli, a difference being determined between ERP associated with common stimuli and ERP associated with the at least one of rare and novel stimuli.

59. The method according to claim 58, wherein, for each of a plurality of pairs of stimuli, the first set of data corresponds to a first time frame ERP₁ immediately after presentation of a first member of one of the pairs of stimuli and a second set corresponds to a second time frame ERP₂ immediately after presentation of a second member of one of the pairs of stimuli, wherein, for each of a first portion of the pairs of stimuli, the second member is the same as the first member and, for each of a second portion of the pairs of stimuli, the second member is different from the first member, ERP₁ and ERP₂ being compared to one another for each of the first and second portions of the pairs of stimuli.

60. The method according to claim 54, wherein the selected features are augmented based on scores from a neurocognitive testing assessment.

61. The method according to claim 54, wherein the selected features are augmented based on scores from a neurobiological marker assessment.

62. The method according to claim 61, wherein the neurobiological marker assessment is one of ApoE and scores derived from clinical historical data.

63. The method according to claim 58, further comprising computing a t-test to assess a significance of a difference between latency periods following presentation of common stimuli and latency periods following the presentation of the at least one of rare and novel stimuli.

64. The method according to claim 63, wherein the latency periods include time periods from 200 to 250 ms and 275 to 350 ms after presentation of the corresponding stimulus.

65. The method according to claim 58, in which stimuli are presented in one of a plurality of sensory modalities and a plurality of stimulus types to a single sensory modality.

66. The method according to claim 59, wherein a preceding time frame immediately after presentation of each of a plurality of individual stimuli is compared to a corresponding time frame immediately after presentation of a succeeding stimulus, with a first portion of comparisons being made for a case where the succeeding stimulus is different from an immediately preceding stimulus and a second portion of the comparisons being made for a case where the succeeding stimulus is the same as the immediately preceding stimulus.

67. The method according to claim 59, wherein ERP₁ and ERP₂ include time intervals approximately 100 to 230 ms after presentation of the corresponding stimulus.

68. The method according to claim 54, wherein the selected features include data corresponding to one of absolute and relative brain wave activity in a delta band.

69. The method according to claim 54, wherein the selected features include data corresponding to one of absolute and relative brain wave activity in a theta band.

70. The method according to claim 54, wherein the selected features include data corresponding to one of absolute and relative brain wave activity in an alpha band.

71. The method according to claim 54, wherein the selected features include data corresponding to one of absolute and relative brain wave activity in a beta band.

72. The method according to claim 54, wherein the selected features include data corresponding to one of absolute and relative brain wave activity in a gamma band.

73. The method according to claim 54, wherein the selected features include data corresponding to one of absolute and relative brain wave activity in an ultrafast frequency band.

74. The method according to claim 73, wherein the ultrafast frequency band includes frequencies ranging from 50 Hz to 1,000 Hz.

75. The method according to claim 54, wherein the selected features include data corresponding to one of brain wave absolute and brain wave relative delta plus theta activity.

76. The method according to claim 54, wherein the selected features include data corresponding to theta spectral activity in posterior and parietal regions and delta spectral power in prefrontal, mesial and frontal regions.

77. The method according to claim 75, wherein the data corresponding to theta spectral activity and delta spectral activity includes a multivariate combination of measurements within these frequency bands based on input from one of the prefrontal and mesial frontal regions.

78. The method according to claim 75, wherein the data corresponding to theta spectral activity and delta spectral activity includes a multivariate combination of measurements within these frequency ranges based on input from one of the frontal, posterior temporal, parietal and occipital regions.

79. The method according to claim 54, wherein the selected features include data corresponding to voltage gradients of activity between and among brain regions.

80. The method according to claim 54, wherein the selected features include one of univariate data features extracted from one of EEG data and MEG data, multivariate data features extracted from one of EEG data and MEG data, Mahalanobis distances extracted from one of EEG data and MEG data and ERP data corresponding to time frames immediately after presentation of a series of stimuli presented in one of a plurality of sensory modalities, a plurality of sensory attributes and a plurality of stimulus attributes within a single sensory modality, including common stimuli and at least one of rare and novel stimuli, a difference being determined between ERP associated with common stimuli and ERP associated with the at least one of rare and novel stimuli.

81. The method according to claim 80, further comprising the step of log transforming the selected features to obtain a normal Gaussian distribution.

82. The method according to claim 54, further comprising the step of removing from the input data one or both of artifacts and data attributable to extracerebral sources.

83. The method according to claim 54, further comprising the step of predicting an expected trajectory of the cognitive state of the individual in a future time period.