INS NYC 2024 Program

Poster	Poster Session 02 Program Schedule 02/15/2024 08:00 am - 09:15 am Room: Shubert Complex (Posters 1-60) Poster Session 02: Aging \| MCI \| Neurodegenerative Disease - PART 1 Final Abstract #17 Lower Learning Ratios Among Amnestic Versus Nonamnestic Subclassifications Anthony Correro II, Medical College of Wisconsin, Milwaukee, United States Laura Glass Umfleet, Medical College of Wisconsin, Milwaukee, United States Russell Bauer, University of Florida, Gainesville, United States Daniel Drane, Emory University, Atlanta, United States Jordan Schwartz, UCLA, Los Angeles, United States David Loring, Emory University, Atlanta, United States Kristen Enriquez, UCLA, Los Angeles, United States Lucia Cavanaugh, UCLA, Los Angeles, United States Joseph Gullett, University of Florida, Gainesville, United States Robert Bilder, UCLA, Los Angeles, United States Category: Assessment/Psychometrics/Methods (Adult) Keyword 1: learning Keyword 2: mild cognitive impairment Keyword 3: dementia - Alzheimer's disease Objective: The Learning Ratio (LR) reflects how much information is learned by the final learning trial relative to the amount of information still-to-be-learned after the first trial. Recent findings suggest lower LRs are associated with poorer delayed recall scores, and LRs outperform other learning slope metrics in differentiating normal cognition, mild cognitive impairment (MCI), and Alzheimer’s disease (AD). This study compares LRs in varied diagnostic groups and memory profiles. We hypothesized that amnestic participants would have lower LRs relative to nonamnestic participants. Participants and Methods: Participant data were extracted from the National Neuropsychology Network data repository. ICD-10 codes were used to determine MCI, dementia, or unspecified neurocognitive disorder diagnoses. Those groups were subclassified as amnestic if delayed recall on any memory test (HVLT-R, Logical Memory, Verbal Paired Associates, CVLT-3) was ≥1.5 SD below the normative mean (i.e., Petersen criteria). The inclusion criteria were age of 55+ years, completion of HVLT-R and WMS-IV Logical Memory (since those tests had the most datapoints), and one of the three target ICD-10 diagnoses. The nonamnestic dementia group was excluded due to small cell size (n=2). The study involved 270 participants (M_age=72.83 years, SD=6.64, 55.19% female; 10.00% non-White; 1.48% Hispanic). A subset of the total sample (n=112) was selected for secondary analyses to resolve problems with circularity (i.e., using the HVLT-R to determine some subclassifications and to calculate LRs). Those participants were categorized as amnestic if their Logical Memory II scaled score was <6. Results: The first groupings (i.e., amnestic or nonamnestic MCI, amnestic dementia, amnestic or nonamnestic unspecified) differed in binary gender distribution (Χ²(4)=10.02, p=.04), so gender was entered into hierarchical multiple regression models first followed by group. Results indicated that group predicted LRs (R²=.12, F(2, 267)=18.42, p<.001, Β_gender=.09 (ns), Β_group=.33), and group added significantly to the prediction model (ΔR²=.11, p<.001). We used a multiple regression with group dummy coded to determine within-group differences. Amnestic groups consistently exhibited lower LRs than nonamnestic groups (B range=.31-.32), but amnestic MCI closely resembled amnestic dementia and amnestic unspecified groups (B range=-.08-.07). For secondary analyses, we computed a pooled variance estimate t-test given unequal sample sizes for the amnestic and nonamnestic groups. Again, the amnestic group had significantly lower LRs (t(110)=-6.16, p<.001, Cohen’s d=-1.19). Conclusions: Amnestic groups consistently had lower LRs relative to nonamnestic groups. LRs were not different between amnestic groups. Prior work suggests the HVLT-R LR differentiates normal cognition from MCI and normal cognition from AD but not MCI from AD. Our results generally align with the notion that HVLT-LR does not distinguish MCI from dementia. We add to the literature by examining MCI subtypes and showed nonamnestic MCI had larger LRs than amnestic dementia. Amnestic individuals are thought to face greater risk for AD compared to other neurodegenerative diseases. Our results may support indications that LRs are more useful than other learning slope scores in staging along the AD continuum.

Poster

02/15/2024
08:00 am - 09:15 am
Room: Shubert Complex (Posters 1-60)

Poster Session 02: Aging | MCI | Neurodegenerative Disease - PART 1

Final Abstract #17

Lower Learning Ratios Among Amnestic Versus Nonamnestic Subclassifications

Anthony Correro II, Medical College of Wisconsin, Milwaukee, United States
Laura Glass Umfleet, Medical College of Wisconsin, Milwaukee, United States
Russell Bauer, University of Florida, Gainesville, United States
Daniel Drane, Emory University, Atlanta, United States
Jordan Schwartz, UCLA, Los Angeles, United States
David Loring, Emory University, Atlanta, United States
Kristen Enriquez, UCLA, Los Angeles, United States
Lucia Cavanaugh, UCLA, Los Angeles, United States
Joseph Gullett, University of Florida, Gainesville, United States
Robert Bilder, UCLA, Los Angeles, United States

Category: Assessment/Psychometrics/Methods (Adult)

Keyword 1: learning
Keyword 2: mild cognitive impairment
Keyword 3: dementia - Alzheimer's disease

Objective:

The Learning Ratio (LR) reflects how much information is learned by the final learning trial relative to the amount of information still-to-be-learned after the first trial. Recent findings suggest lower LRs are associated with poorer delayed recall scores, and LRs outperform other learning slope metrics in differentiating normal cognition, mild cognitive impairment (MCI), and Alzheimer’s disease (AD). This study compares LRs in varied diagnostic groups and memory profiles. We hypothesized that amnestic participants would have lower LRs relative to nonamnestic participants.

Participants and Methods:

Participant data were extracted from the National Neuropsychology Network data repository. ICD-10 codes were used to determine MCI, dementia, or unspecified neurocognitive disorder diagnoses. Those groups were subclassified as amnestic if delayed recall on any memory test (HVLT-R, Logical Memory, Verbal Paired Associates, CVLT-3) was ≥1.5 SD below the normative mean (i.e., Petersen criteria). The inclusion criteria were age of 55+ years, completion of HVLT-R and WMS-IV Logical Memory (since those tests had the most datapoints), and one of the three target ICD-10 diagnoses. The nonamnestic dementia group was excluded due to small cell size (n=2). The study involved 270 participants (M_age=72.83 years, SD=6.64, 55.19% female; 10.00% non-White; 1.48% Hispanic). A subset of the total sample (n=112) was selected for secondary analyses to resolve problems with circularity (i.e., using the HVLT-R to determine some subclassifications and to calculate LRs). Those participants were categorized as amnestic if their Logical Memory II scaled score was <6.

Results:

The first groupings (i.e., amnestic or nonamnestic MCI, amnestic dementia, amnestic or nonamnestic unspecified) differed in binary gender distribution (Χ²(4)=10.02, p=.04), so gender was entered into hierarchical multiple regression models first followed by group. Results indicated that group predicted LRs (R²=.12, F(2, 267)=18.42, p<.001, Β_gender=.09 (ns), Β_group=.33), and group added significantly to the prediction model (ΔR²=.11, p<.001). We used a multiple regression with group dummy coded to determine within-group differences. Amnestic groups consistently exhibited lower LRs than nonamnestic groups (B range=.31-.32), but amnestic MCI closely resembled amnestic dementia and amnestic unspecified groups (B range=-.08-.07). For secondary analyses, we computed a pooled variance estimate t-test given unequal sample sizes for the amnestic and nonamnestic groups. Again, the amnestic group had significantly lower LRs (t(110)=-6.16, p<.001, Cohen’s d=-1.19).

Conclusions:

Amnestic groups consistently had lower LRs relative to nonamnestic groups. LRs were not different between amnestic groups. Prior work suggests the HVLT-R LR differentiates normal cognition from MCI and normal cognition from AD but not MCI from AD. Our results generally align with the notion that HVLT-LR does not distinguish MCI from dementia. We add to the literature by examining MCI subtypes and showed nonamnestic MCI had larger LRs than amnestic dementia. Amnestic individuals are thought to face greater risk for AD compared to other neurodegenerative diseases. Our results may support indications that LRs are more useful than other learning slope scores in staging along the AD continuum.