Differentiating alzheimer's disease from vascular dementia with EEG functional neuroimaging

K. Kilborn, J. Price, Z. Tieges, B.A. Conway, Susil Stephen, A. Hughes, G.S. McLean

    Research output: Contribution to journalConference Contribution

    Abstract

    Differential diagnosis of Alzheimer's disease (AD) and vascular dementia (VaD) is challenging because early symptoms are similar despite different underlying neuropathology. In previous research, we found that a computerized associative memory task coupled with high resolution EEG measures differentiated mild AD from healthy controls (85% sensitivity, 93% specificity). We extend this functional imaging technique to identify differential diagnostic markers in AD and VaD.
    Participants included untreated patients newly diagnosed with probable AD (N = 8), VaD (N = 8), and controls (N = 8). Patients were matched for MMSE scores, and all participants were matched for age and gender. Each participant completed a 20minute computerized memory task. Stimuli consisted of simple visual images paired with associatively related auditory words. Participants were required to respond Old or New by button press on each trial. EEG (128 channel) was recorded continuously during the task. We examined ERP components including the N100, P100, N200, P300, P400, P600, a difference wave at 700 ms, and behavioral responses (d'). Each variable was assessed individually using binary logistic regression, and sensitivity and specificity profiles were constructed. Next, the best classifiers were entered into regression analyses in pairs, and then in groups of three.
    Good discrimination was achieved by several combinations of 2 and 3 variables, with sensitivity ranging from 62% to 87.5%, and specificity from 75% to 87.5%. The best classification between AD and VaD groups was achieved by a centro-parietal N200 amplitude difference combined with the 700ms difference wave (87.5% sensitivity and 87.5% specificity). These same ERP variables also discriminated between VaD patients and controls (75% sensitivity and 75% specificity). The behavioral measure (d' statistic) discriminated between controls and the two patient groups, but not between AD and VaD groups.
    Despite similarity in early symptoms in AD and VaD, differences in the underlying neuropathology may produce a reliably different signature in the EEG signal. We found that specific combinations of ERP effects generated during a controlled memory test provide good sensitivity/specificity with respect to AD, VaD, and matched controls. Such functional biomarkers may offer a clinically useful aid in differential diagnosis of common forms of dementia.

    Fingerprint

    Functional neuroimaging
    Functional Neuroimaging
    Vascular Dementia
    Electroencephalography
    Alzheimer Disease
    Enterprise resource planning
    Sensitivity and Specificity
    Data storage equipment
    Biomarkers
    Differential Diagnosis
    Logistics
    Classifiers
    Statistics
    Imaging techniques

    Keywords

    • alzheimer's disease
    • vascular dementia
    • EEG functional neuroimaging

    Cite this

    Kilborn, K. ; Price, J. ; Tieges, Z. ; Conway, B.A. ; Stephen, Susil ; Hughes, A. ; McLean, G.S. / Differentiating alzheimer's disease from vascular dementia with EEG functional neuroimaging. In: Alzheimer's and Dementia. 2009 ; Vol. 5, No. 4. pp. e24-e24.
    @article{9268d087e19648e49974b49a0a493976,
    title = "Differentiating alzheimer's disease from vascular dementia with EEG functional neuroimaging",
    abstract = "Differential diagnosis of Alzheimer's disease (AD) and vascular dementia (VaD) is challenging because early symptoms are similar despite different underlying neuropathology. In previous research, we found that a computerized associative memory task coupled with high resolution EEG measures differentiated mild AD from healthy controls (85{\%} sensitivity, 93{\%} specificity). We extend this functional imaging technique to identify differential diagnostic markers in AD and VaD. Participants included untreated patients newly diagnosed with probable AD (N = 8), VaD (N = 8), and controls (N = 8). Patients were matched for MMSE scores, and all participants were matched for age and gender. Each participant completed a 20minute computerized memory task. Stimuli consisted of simple visual images paired with associatively related auditory words. Participants were required to respond Old or New by button press on each trial. EEG (128 channel) was recorded continuously during the task. We examined ERP components including the N100, P100, N200, P300, P400, P600, a difference wave at 700 ms, and behavioral responses (d'). Each variable was assessed individually using binary logistic regression, and sensitivity and specificity profiles were constructed. Next, the best classifiers were entered into regression analyses in pairs, and then in groups of three.Good discrimination was achieved by several combinations of 2 and 3 variables, with sensitivity ranging from 62{\%} to 87.5{\%}, and specificity from 75{\%} to 87.5{\%}. The best classification between AD and VaD groups was achieved by a centro-parietal N200 amplitude difference combined with the 700ms difference wave (87.5{\%} sensitivity and 87.5{\%} specificity). These same ERP variables also discriminated between VaD patients and controls (75{\%} sensitivity and 75{\%} specificity). The behavioral measure (d' statistic) discriminated between controls and the two patient groups, but not between AD and VaD groups. Despite similarity in early symptoms in AD and VaD, differences in the underlying neuropathology may produce a reliably different signature in the EEG signal. We found that specific combinations of ERP effects generated during a controlled memory test provide good sensitivity/specificity with respect to AD, VaD, and matched controls. Such functional biomarkers may offer a clinically useful aid in differential diagnosis of common forms of dementia.",
    keywords = "alzheimer's disease, vascular dementia, EEG functional neuroimaging",
    author = "K. Kilborn and J. Price and Z. Tieges and B.A. Conway and Susil Stephen and A. Hughes and G.S. McLean",
    year = "2009",
    doi = "10.1016/j.jalz.2009.07.097",
    language = "English",
    volume = "5",
    pages = "e24--e24",
    journal = "Alzheimer's and Dementia",
    issn = "1552-5260",
    number = "4",

    }

    Differentiating alzheimer's disease from vascular dementia with EEG functional neuroimaging. / Kilborn, K.; Price, J.; Tieges, Z.; Conway, B.A.; Stephen, Susil; Hughes, A.; McLean, G.S.

    In: Alzheimer's and Dementia, Vol. 5, No. 4, 2009, p. e24-e24.

    Research output: Contribution to journalConference Contribution

    TY - JOUR

    T1 - Differentiating alzheimer's disease from vascular dementia with EEG functional neuroimaging

    AU - Kilborn, K.

    AU - Price, J.

    AU - Tieges, Z.

    AU - Conway, B.A.

    AU - Stephen, Susil

    AU - Hughes, A.

    AU - McLean, G.S.

    PY - 2009

    Y1 - 2009

    N2 - Differential diagnosis of Alzheimer's disease (AD) and vascular dementia (VaD) is challenging because early symptoms are similar despite different underlying neuropathology. In previous research, we found that a computerized associative memory task coupled with high resolution EEG measures differentiated mild AD from healthy controls (85% sensitivity, 93% specificity). We extend this functional imaging technique to identify differential diagnostic markers in AD and VaD. Participants included untreated patients newly diagnosed with probable AD (N = 8), VaD (N = 8), and controls (N = 8). Patients were matched for MMSE scores, and all participants were matched for age and gender. Each participant completed a 20minute computerized memory task. Stimuli consisted of simple visual images paired with associatively related auditory words. Participants were required to respond Old or New by button press on each trial. EEG (128 channel) was recorded continuously during the task. We examined ERP components including the N100, P100, N200, P300, P400, P600, a difference wave at 700 ms, and behavioral responses (d'). Each variable was assessed individually using binary logistic regression, and sensitivity and specificity profiles were constructed. Next, the best classifiers were entered into regression analyses in pairs, and then in groups of three.Good discrimination was achieved by several combinations of 2 and 3 variables, with sensitivity ranging from 62% to 87.5%, and specificity from 75% to 87.5%. The best classification between AD and VaD groups was achieved by a centro-parietal N200 amplitude difference combined with the 700ms difference wave (87.5% sensitivity and 87.5% specificity). These same ERP variables also discriminated between VaD patients and controls (75% sensitivity and 75% specificity). The behavioral measure (d' statistic) discriminated between controls and the two patient groups, but not between AD and VaD groups. Despite similarity in early symptoms in AD and VaD, differences in the underlying neuropathology may produce a reliably different signature in the EEG signal. We found that specific combinations of ERP effects generated during a controlled memory test provide good sensitivity/specificity with respect to AD, VaD, and matched controls. Such functional biomarkers may offer a clinically useful aid in differential diagnosis of common forms of dementia.

    AB - Differential diagnosis of Alzheimer's disease (AD) and vascular dementia (VaD) is challenging because early symptoms are similar despite different underlying neuropathology. In previous research, we found that a computerized associative memory task coupled with high resolution EEG measures differentiated mild AD from healthy controls (85% sensitivity, 93% specificity). We extend this functional imaging technique to identify differential diagnostic markers in AD and VaD. Participants included untreated patients newly diagnosed with probable AD (N = 8), VaD (N = 8), and controls (N = 8). Patients were matched for MMSE scores, and all participants were matched for age and gender. Each participant completed a 20minute computerized memory task. Stimuli consisted of simple visual images paired with associatively related auditory words. Participants were required to respond Old or New by button press on each trial. EEG (128 channel) was recorded continuously during the task. We examined ERP components including the N100, P100, N200, P300, P400, P600, a difference wave at 700 ms, and behavioral responses (d'). Each variable was assessed individually using binary logistic regression, and sensitivity and specificity profiles were constructed. Next, the best classifiers were entered into regression analyses in pairs, and then in groups of three.Good discrimination was achieved by several combinations of 2 and 3 variables, with sensitivity ranging from 62% to 87.5%, and specificity from 75% to 87.5%. The best classification between AD and VaD groups was achieved by a centro-parietal N200 amplitude difference combined with the 700ms difference wave (87.5% sensitivity and 87.5% specificity). These same ERP variables also discriminated between VaD patients and controls (75% sensitivity and 75% specificity). The behavioral measure (d' statistic) discriminated between controls and the two patient groups, but not between AD and VaD groups. Despite similarity in early symptoms in AD and VaD, differences in the underlying neuropathology may produce a reliably different signature in the EEG signal. We found that specific combinations of ERP effects generated during a controlled memory test provide good sensitivity/specificity with respect to AD, VaD, and matched controls. Such functional biomarkers may offer a clinically useful aid in differential diagnosis of common forms of dementia.

    KW - alzheimer's disease

    KW - vascular dementia

    KW - EEG functional neuroimaging

    UR - http://day.alzheimersanddementia.com/home

    U2 - 10.1016/j.jalz.2009.07.097

    DO - 10.1016/j.jalz.2009.07.097

    M3 - Conference Contribution

    VL - 5

    SP - e24-e24

    JO - Alzheimer's and Dementia

    T2 - Alzheimer's and Dementia

    JF - Alzheimer's and Dementia

    SN - 1552-5260

    IS - 4

    ER -