Classification of Patients with Mild Cognitive Impairment vs. Normal Controls based on Experimental and Conventional Standardized Measures of Processing Speed and Working Memory.

1

• Classification of Patients with Mild Cognitive
  Impairment vs. Normal Controls based on
  Experimental and Conventional Standardized
  Measures of Processing Speed and Working Memory.
• Erin Schlicting1, Stephen Correia2, Paul Malloy2,
  Stephen Salloway2
• 1Department of Psychology, University of Rhode
  Island 2Department of Psychiatry and Human
  Behavior Brown Medical School Butler Hospital,
  Providence, RI

Table 2 Cognitive Measures
Table 3 Variables selected from the MANOVA
Objective
Standardized Measures Standardized Measures Standardized Measures
Processing Speed ES ES Working Memory ES Executive Function ES
Symbol Digit 0.48 0.48 Letter-Number Sequencing 0.11 COWA 0.07
Trails A 0.21 0.21 Digit Span - Total 0.22 Trails B - Trails A 0.10
Grooved Pegboard     Digit Span - Backward 0.20 Trails B 0.21
Dominant 0.17 0.17 Spatial Span -Total 0.17
Non-Dominant 0.20 0.20 Spatial Span - Backward 0.28    

Experimental Measures Experimental Measures Experimental Measures        
Processing Speed Processing Speed ES Working Memory ES Executive Function ES
0-Back Rxn Time 0-Back Rxn Time 0.14 SOPT Abstract Designs 0.37 N/A N/A
1-Back Rxn Time 1-Back Rxn Time 0.10 SOPT Abstract Words 0.08
2-Back Rxn Time 2-Back Rxn Time 0.09 1-Back Correct Responses 0.21
    2-Back Correct Responses 0.24
      3-Back Correct Responses 0.40    
Effect sizes determined from values for partial ?2 .06-.12 Medium Effect gt .12 Large Effect Effect sizes determined from values for partial ?2 .06-.12 Medium Effect gt .12 Large Effect Effect sizes determined from values for partial ?2 .06-.12 Medium Effect gt .12 Large Effect Effect sizes determined from values for partial ?2 .06-.12 Medium Effect gt .12 Large Effect Effect sizes determined from values for partial ?2 .06-.12 Medium Effect gt .12 Large Effect Effect sizes determined from values for partial ?2 .06-.12 Medium Effect gt .12 Large Effect Effect sizes determined from values for partial ?2 .06-.12 Medium Effect gt .12 Large Effect
     

• To determine if N-back and Self-Ordered Pointing
  Tasks (SOPT) improves classification of patients
  with amnestic mild cognitive impairment (MCI) vs.
  cognitively normal controls (NC) over and above
  standardized tests of processing speed and
  working memory.

Measure Processing Speed Working Memory Executive Function
Standardized Symbol Digit Digit Span DRS IP Subtest
Trails A Spatial Span DRS AC Subtest
Grooved Pegboard Letter-Number Seq. Trails B
  Working Memory Index COWA
      Trails B-Trails A
Experimental 0-back Nback Task 1, 2, 3 Nback tasks N/A
(correct responses) (correct responses)
0, 1, 2, 3 Nback SOPT Designs
  (reaction time) and Words  
Experimental measures of executive function were not examined in this study Experimental measures of executive function were not examined in this study Experimental measures of executive function were not examined in this study Experimental measures of executive function were not examined in this study
Background

• Working memory (WM) and processing speed (PS) are
  often impaired in patients with amnestic MCI1.
• Patients with MCI and deficits in WM and PS are
  at greater risk for conversion to dementia2,3.
• Executive impairment (including WM and PS) adds
  to functional disability in Alzheimers
  disease4,5.
• Experimental tests of WM and PS may improve
  discrimination between MCI and NC when combined
  with standardized WM and PS.
• Identifying patients at greatest risk for
  dementia is important for implementing early
  interventions.

Analysis

• Correlations among cognitive variables within
  groups were examined to evaluate for possible
  collinearity.
• MANOVA procedure was used to identify significant
  differences between MCIs and NCs on
  standardized and experimental tests of WM and PS.
  
• Three DFA procedures were conducted, one using
  the standardized tests of WM, PS, and EF, a
  second using experimental test of WM and PS, and
  a third DFA combining measures retained from the
  two prior DFAs.
• The effect sizes for the three DFAs were
  examined to determine if combining standardized
  and experimental measures improved classification
  accuracy (MCI vs. NC).

Table 4 Discriminant Function Analyses
DFA Procedure Measures Retained Effect Size
1. Standardized Measures   0.52
Processing Speed Symbol Digit 0.71
Working Memory Spatial Span - Backward 0.30
2. Experimental Measures   0.60
Processing Speed NS NS
Working Memory SOPT Abstract Designs 0.59
3-Back Correct Responses 0.62
3. Combined Measures   0.72
Processing Speed Symbol Digit 0.25
Working Memory Spatial Span - Backward 0.15
SOPT Abstract Designs 0.35
3-Back Correct Responses 0.25
Effect sizes for the DFAs and measures retained in each discriminant function Effect sizes for the DFAs and measures retained in each discriminant function Effect sizes for the DFAs and measures retained in each discriminant function
Participants

• Participants were 26 patients with amnestic MCI
  (Petersen criteria6) and 20 age matched cognitive
  normal elderly controls (Table 1).
• All participants undertook a battery of cognitive
  tests including standardized and experimental
  measures of executive function, WM, and PS (Table
  2).

Results

• The groups did not significantly by age or
  education (Table 1).
• No measures were excluded due to collinearity
  (all r lt .65).
• MANOVA revealed a significant main effect for
  group (NC vs. MCI) plt.001.
• Measures from the MANOVA with medium effect sizes
  (ES, partial ?2 gt.06 were selected for each DFA
  (Table 3).
• Each DFA produced one significant function
• Standardized measures (Wilks ? .48 plt.001)
• Experimental measures (Wilks ? .40, plt.001)
• Combined measures (Wilks ? .28, plt.001)
• Table 4 lists the effect size for each DFA and
  the tests retained with corresponding effect
  sizes (1-Wilks ?).
• The combination of standardized and experimental
  tests of PS and WM was the strongest
  discriminator of MCI from NC.

Table 1 Participant Characteristics
Conclusions
  Normal Normal Mild Cognitive Mild Cognitive Significant
  Controls Controls Impairment Impairment Differences
  (n 20 55 Female) (n 20 55 Female) (n 26 42 Female) (n 26 42 Female)  
Variable Mean SD Mean SD p Value
Age 71.5 10.0 75.6 7.6 0.144
Education 14.9 3.0 14.2 3.4 0.940

• The results are consistent with prior findings of
  deficits in PS and WM in amnestic MCI1,3.
• Performance on the SOPT and 3-Back tests improved
  discrimination of MCI from NC over and above
  standardized tests of PS and WM.
• WM performance is impaired in MCI and may help
  discriminating patients in the early stages of
  Alzheimers disease from controls7.

References 1.Nordlund, A. et al. The Goteborg
MCI study mild cognitive impairment is a
heterogeneous condition. J Neurol Neurosurg
Psychiatry 76, 1485-90 (2005). 2.Albert, M.S.,
Moss, M.B., Tanzi, R. Jones, K. Preclinical
prediction of AD using neuropsychological tests.
J Int Neuropsychol Soc 7, 631-9 (2001). 3.Tabert,
M.H. et al. Neuropsychological prediction of
conversion to Alzheimer disease in patients with
mild cognitive impairment. Arch Gen Psychiatry
63, 916-24 (2006). 4.Boyle, P.A. et al. Executive
dysfunction and apathy predict functional
impairment in Alzheimer disease. Am J Geriatr
Psychiatry 11, 214-21 (2003). 5.Cahn-Weiner,
D.A., Boyle, P.A. Malloy, P.F. Tests of
executive function predict instrumental
activities of daily living in community-dwelling
older individuals. Appl Neuropsychol 9, 187-91
(2002). 6.Petersen, R.C. et al. Mild cognitive
impairment clinical characterization and
outcome. Arch Neurol 56, 303-8 (1999). 7.
Germano, C. Kinsella, G.J. Working memory and
learning in early Alzheimer's disease.
Neuropsychology review 15, 1-10 (2005).
Acknowledgments Support from NIA ZAG1 FAS-5
(T32) Alzheimers Association NIRG-03-6195 NIMH
K08MH01487W The Human Brain Project (NIBIB
NIMH) Ittleson Fund at Brown P20 NCRR15578-01
Center for Translational Brain Research at Brown.