Table 1.
Socio-demographic data in normal controls, schizophrenia patients with and without deficit schizophrenia and patients with amnestic mild cognitive impairment (aMCI).
Table 2.
Correlation matrix between the Consortium to Establish a Registry for Alzheimer’s disease test results and age and education in healthy controls.
Table 3.
Results of multivariate general linear model analyses with the Consortium to Establish a Registry for Alzheimer’s disease (CERAD) tests results as dependent variables and diagnosis as primary explanatory variable, while adjusting for age, sex and education.
Shown are the model-generated marginal means (SE) of the z-scores of all CERAD tests.
Fig 1.
Mean (SE) z scores of Consortium to Establish a Registry for Alzheimer’s disease tests in normal controls, nondeficit and deficit schizophrenia patients and patients with amnestic mild cognitive impairment (MCI).
VFT24: verbal fluency test; BNT: Modified Boston Naming Test; BNTh: naming of highly frequency objects; BNTm: naming of medium frequency objects; BNTl: naming of low frequency objects; BNTtotal: BNT total score. MMSE: Mini Mental State Examination.
Fig 2.
Mean (SE) z scores of Consortium to Establish a Registry for Alzheimer’s disease tests in normal controls, nondeficit and deficit schizophrenia patients and patients with amnestic mild cognitive impairment (MCI).
WLcorrect1: Trial 1 (WLM correct 1); WLcorrect2: Trial 2 (WLM correct 2); WLcorrect3: Trial 3 (WLM correct 3); WLM: Word List Memory, total score.
Fig 3.
Mean (SE) z scores of Consortium to Establish a Registry for Alzheimer’s disease tests in normal controls, nondeficit and deficit schizophrenia patients and patients with amnestic mild cognitive impairment (MCI).
TrueWLRecall: Word List Recall, Delayed, true recall; FalseWLRecall: Word List Recall, Delayed, false recall (WL False Recall); CorrectWLRecognition: WL Recognition Correct Yes response; NoWLRecognition: Word List Recognition Correct No response; WLRecognition: sum of Correctand No WLRecognition.
Table 4.
Results of binary logistic regression analyses with amnestic mild cognitive impairment (aMCI) versus deficit or non-deficit schizophrenia as classes and 7 Consortium to Establish a Registry for Alzheimer’s disease (CERAD) tests as explanatory variables.
Table 5.
Results of Linear SVM and Random Forest Models.
Table 6.
Results of Multilayer Perceptron Neural Network analysis with 7 Consortium to Establish a Registry for Alzheimer’s disease (CERAD) tests.
Fig 4.
Neural network importance chart showing the relative and normalized importances of the Consortium to Establish a Registry for Alzheimer’s disease test scores as input variables predicting amnestic mild cognitive impairment versus deficit schizophrenia (output variables).
BNTtotal: Boston naming Test, total score; WLM: Word List Memory; WLFalse Recall: Word List Recall, Delayed, false recall; MMSE: Mini Mental State Examination; WLTRueRecall: Word List Recall, Delayed, true recall; VFT: Verbal Fluency Test; WLRecognition: Word List Recognition, total score.
Fig 5.
Neural network importance chart showing the relative and normalized importances of the Consortium to Establish a Registry for Alzheimer’s disease test scores as input variables predicting amnestic mild cognitive impairment versus nondeficit schizophrenia (output variables).
BNTtotal: Boston naming Test, total score; WLM: Word List Memory; WLTRueRecall: Word List Recall, Delayed, true recall; WLRecognition: Word List Recognition, total score; WLFalse Recall: Word List Recall, Delayed, false recall; VFT: Verbal Fluency Test; MMSE: Mini Mental State Examination.