MCI Definitions in EAS

Mild cognitive impairment (MCI) is a condition in which people have more memory or thinking problems than other people their age. By clinical definition, MCI is a transitional stage between normal cognitive performance (often age and demographically adjusted) and dementia. The concept of MCI is strictly applicable only to people who do not have dementia. Two main criteria used for MCI are Petersen and Jak-Bondi.

Petersen

• An earlier (Petersen 1999) definition that relies heavily on a participant’s subjective symptoms combined with a broad objective testing using a threshold.
• Based on objective cognitive tests, subjective cognitive concerns (self or informant), and functional.
• It also defines subtype of MCI: amnestic (aMCI) and non-amnestic (naMCI) status.

Jak-Bondi

• A more recent (Bondi et al 2014, Jack et al 2016) definition using actuarial method that relies on patterns across multiple tests.
• Based only on objective cognitive tests and functional status.

Definitions Used in EAS

• The Petersen definition of MCI was used in old EAS (prior to 2017).
• The Jak-Bondi definition was adopted in new EAS (post 2017).
• We now provide both definitions in the EAS database throughout the time frames.

Objective Cognitive Tests

EAS collects 2 objective cognitive tests accross 5 domains.

Domain	Old EAS	New EAS
Memory	FCSRT Free recall Logical Memory (Story A, Immediate)*	FCSRT Free recall Benson Complex Figure (Delayed)
Attention	Digit span* Trail Making Test Part A	Number Span Trail Making Test Part A
Executive function	Trail Making Test Part B Phonemic Verbal Fluency (FS)	Trail Making Test Part B Phonemic Verbal Fluency (FS)
Language	Boston Naming Test (BNT, total)* Category Fluency (animal, vegetable)	Multilingual Naming Test (MINT, total) Category Fluency (animal, vegetable)
Visuospatial	WAIS III Block Design Digit symbol	WAIS III Block Design Benson Complex Figure Copy

• Tests marked with * have not been collected post 2017 (Logical Memory since 2023).
• In the new EAS, Trail Making status and notes allow us to determine if missing values on Trail Making, mostly Part B, were due to cognitive reasons; if so, missing values were assigned the maximum limit value.

Cognitive Norms

Determination of whether a cognitive test score is impaired is done by comparing it with age, sex, education adjusted mean value in a cognitively normal population (further race/ethnic adjusted can also be considered). We used local cognitive norms defined as cognitively unimpaired at baseline (wave=1) for both old and new EAS: not demented, CDR=0. Linear regression models for cognitive test scores on age, sex, and education were fit using the normative data.

Z-Scores and Cut-Off Values

In EAS data, standardized Z-scores for the cognitive tests can be calculated using expected age, sex and education adjusted mean score and variation from the norms.

Trail Making times are reversed (negative) so that higher cognitive test scores in all tests indicate better performance.

In the Petersen definition, a cognitive test score is considered impaired if more than 1.5 SD below the mean (i.e., Z < −1.5).

In JB definition, a cognitive test score is considered impaired if more than 1 SD below the mean (i.e., Z < −1).

Data Comparison

Old EAS Baseline Norm Data (n=1221)

Variable	Mean	StdDev	Median	Minimum	Q1	Q3	Maximum
age	77.78	5.3	77.05	66.42	73.48	80.93	100.43
Educyrs	13.73	3.48	13	2	12	16	24
female(%)	61.3%
FreeRecall	31.38	5.77	32	4	28	35	45
StoryA	10.94	3.87	11	0	8	14	22
digitspan	14.04	3.69	14	2	12	16	27
TRAILA	58.51	25.28	52	16.7	42	69	238.76
TRAILB	141	70.2	120	9.93	88	174.06	300
FS	24.56	8.91	24	3	18	30	53
BNT	12.19	2.41	13	3	11	14	15
CATTotal	25.59	6.63	25	8	21	30	51
Blockraw	21.41	9.12	21	0	16	27	61
digitsymbol	41.62	14.38	41	0	32	51	85

New EAS Baseline Norm Data (n=264)

Variable	Mean	StdDev	Median	Minimum	Q1	Q3	Maximum
Age	73.91	6.19	73.95	60.1	70.65	77.65	97.4
Educyrs	15.77	3.45	16	7	13	18	29
female(%)	66.7%
MocaScore	23.9	3.52	25	14	22	27	30
FCSRFree	33.42	4.79	34	19	30	37	45
BensonScoreD	10.19	3.2	10	0	8	12	16
FS	27.82	8.29	27	9	22	34	53
TRAILA	45.7	20.51	41.3	15.9	32	52.8	150
TRAILB	130.18	75.74	102.5	41	78.5	159	300
NumberSpan	13.06	3.8	13	5	10	15	25
CATTotal	30.11	7.78	30	12	24	35	51
MintTotal	27.47	3.86	28	10	25	30	32
BlockRaw	24.86	9.91	25	9	16	30	61
BensonScoreI	15.14	1.14	16	12	15	16	16
Digitsymbol	48.78	13.95	48	10	40	57	106
VerbatimI	17.94	6.37	18	2	13	23	44
ParaphraseI	14.21	4.15	14	3	11	18	23
VerbatimD	15.1	6.13	15	1	11	19	33
ParaphraseD	13.13	4.18	13	1	10.5	16	22

Harmonization of Cognitive Tests

Three cognitive tests used in old EAS were replaced by new tests post 2017:

• Logical Memory was replaced by Craft Story 21.
• Digit Span was replaced by Number Span (nearly identical, UDS v3 Number Span is a nonproprietary forward-and-backward digit-span measure that closely parallels the structure of WAIS-III Digit Span but uses different number sequences).
• BNT was replaced by MINT.

The NACC Crosswalk study provided equipercentile conversion from new tests to the corresponding old tests. When determining MCI status at visits post 2017 for participants enrolled before 2017, we continue to use old tests, applying conversion from the corresponding new tests.

Functional Component

Functional status is determined from 4 IADL items (ADCS since 2023):

• Telephone use
• Transportation
• Medication
• Finance

Functional impairment is defined by impaired in all the 4 tasks.

Subjective Cognitive Concerns (SCC)

Self-reported subjective cognitive concerns (have SCC or not) are measured by CERAD1 (multiple items) and GDS10 in old EAS. In the New EAS, a questionnaire that assesses perceived cognitive problems and worries is collected by phone screening.

Petersen MCI

A participant is considered to have mild cognitive impairment (MCI=1) if all the following 3 conditions are met:

1. Have SCC, and
2. Impaired in at least one of the 10 cognitive tests, and
3. Not functionally impaired.

Petersen MCI Subtypes

MCI is categorized as amnestic (aMCI), where memory loss is the main symptom, or non-amnestic (naMCI), where memory is intact but other domains (attention, executive function, language, visual-spacial) are impaired.

• Criteria for aMCI: (1) Have SCC, and (2) impaired in at least one of the 2 memory tests, and (3) not functionally impaired.
• Criteria for naMCI: (1) Have SCC, and (2) impaired in at least one of the 8 non-memory tests and not impaired in either memory test, and (3) not functionally impaired.

Further Classification of MCI Subtypes Based on Number of Domains Involved

aMCI

• Single-domain aMCI: Only memory is impaired.
• Multiple-domain aMCI: Memory is impaired alongside impairment in one or more other domains.

naMCI

• Single-domain naMCI: Impairment is limited to a single non-memory domain.
• Multiple-domain naMCI: Impairments occur in two or more non-memory domains.

JB MCI

A participant is considered to have mild cognitive impairment (MCI=1) if any one of the following 3 conditions are met:

1. Impaired scores on both cognitive tests within at least one cognitive domain; or
2. At least one impaired score in at least three of the five cognitive domains; or
3. Functionally impaired.

Summary: MCI Variables in EAS

Note that MCI is only applicable to those not diagnosed as dementia. For dementia cases, all MCI variables are assigned value of 9. Types/subtypes of MCI are only defined for Petersen definition.

MCI Variable	Description
JMCI	JB definition of MCI
PMCI	Petersen definition of MCI
aMCI	Amnestic MCI (Petersen)
naMCI	Non-amnestic MCI (Petersen)
aMCI cat	Further categorization of aMCI
naMCI cat	Further categorization of naMCI

Codes for JMCI, PMCI, aMCI, and naMCI:

• 0 = not impaired
• 1 = impaired
• 9 = dementia

Codes for aMCI cat and naMCI cat:

• 0 = not impaired
• 1 = single-domain impaired
• 2 = multiple-domain impaired
• 9 = dementia

References

Petersen

• Petersen, R. C. (2004). Mild cognitive impairment as a diagnostic entity. Journal of Internal Medicine, 256(3), 183–194.
• Winblad, B., Palmer, K., Kivipelto, M., Jelic, V., Fratiglioni, L., Wahlund, L-O., Nordberg, A., Bäckman, L., Albert, M., Almkvist, O., Arai, H., Basun, H., Blennow, K., De Leon, M., DeCarli, C., Erkinjuntti, T., Giacobini, E., Graff, C., Hardy, J., Petersen, R. C. (2004). Mild cognitive impairment-beyond controversies, towards a consensus: Report of the international working group on mild cognitive impairment. Journal of Internal Medicine, 256, 240–246.

Jak-Bondi

• Bondi, M. W., Edmonds, E. C., Jak, A. J., et al. (2014). Neuropsychological criteria for mild cognitive impairment improves diagnostic precision, biomarker associations, and progression rates. Journal of Alzheimer’s Disease, 42(1), 275–289.
• Jak, AJ, Preis, SR, Beiser, AS, Seshadri, S, Wolf, PA., Bondi, MW, Au, R. (2016). Neuropsychological criteria for mild cognitive impairment and dementia risk in the Framingham Heart Study. J Int Neuropsychol Soc, 22(9), 937-943.

MCI in EAS and Comparison

• Katz MJ, Wang C, Nester CO, Derby CA, Zimmerman ME, Lipton RB, Sliwinski MJ, Rabin LA. T-MoCA: A valid phone screen for cognitive impairment in diverse community samples. Alzheimers Dement (Amst). 2021 Feb 5;13(1):e12144. doi: 10.1002/dad2.12144. PMID: 33598528; PMCID: PMC7864219. (JB MCI in new EAS)
• Chang, K. H., Wang, C., Qin, J., Katz, M. J., Byrd, D. A., Lipton, R. B., Rabin, L. A. (2025). Comparing the predictive validity of four MCI definitions for incident dementia in demographically diverse community-dwelling individuals: Results from the Einstein Aging Study (EAS). Journal of the International Neuropsychological Society : JINS, 31(1), 32–41. https://doi.org/10.1017/S1355617724000729 (comparison of 4 MCI definitions using old EAS data)