Academic Journal
View in EDS

Development and External Validation of a Machine Learning Model for Classification of Mild Cognitive Impairment and Dementia Using Clinical Data.

Bibliographic Details
Title: Development and External Validation of a Machine Learning Model for Classification of Mild Cognitive Impairment and Dementia Using Clinical Data.
Authors: Kannenieks D; Faculty of Medicine, Riga Stradins University, LV-1007 Riga, Latvia., Priede Z; Department of Neurology and Neurosurgery, Riga Stradins University, LV-1002 Riga, Latvia., Millers A; Department of Neurology and Neurosurgery, Riga Stradins University, LV-1002 Riga, Latvia., Velins KK; Department of Mathematics, Imperial College London, London SW7 2AZ, UK.
Source: Medicina (Kaunas, Lithuania) [Medicina (Kaunas)] 2026 Jul 14; Vol. 62 (7). Date of Electronic Publication: 2026 Jul 14.
Publication Type: Journal Article
Language: English
Journal Info: Publisher: MDPI Country of Publication: Switzerland NLM ID: 9425208 Publication Model: Electronic Cited Medium: Internet ISSN: 1648-9144 (Electronic) Linking ISSN: 1010660X NLM ISO Abbreviation: Medicina (Kaunas) Subsets: MEDLINE
Imprint Name(s): Publication: 2018- : Basel, Switzerland : MDPI
Original Publication: Kaunas : Lietuvos gydytojų sąjunga
MeSH Terms: Cognitive Dysfunction*/classification , Cognitive Dysfunction*/diagnosis , Machine Learning*/standards , Dementia*/classification , Dementia*/diagnosis, Humans ; Aged ; Retrospective Studies ; Male ; Female ; Boosting Machine Learning Algorithms ; Aged, 80 and over ; Cohort Studies ; Classification Algorithms ; Predictive Learning Models
Abstract: Background and Objectives: As society ages, the number of patients with cognitive impairment is increasing. Machine learning methods that use structured clinical and cognitive-assessment data during routine diagnostic work-up may support and monitor structured classifications of cognitive status. This kind of approach can improve early screening, reduce physicians' workload and develop greater support for personalized treatment. To develop an XGBoost-based machine learning model using the National Alzheimer's Coordinating Center (NACC) dataset and to evaluate the model's precision with clinician-assigned diagnosis in a Latvian retrospective cohort study. Materials and Methods: The research was designed as a retrospective external validation cohort study that used two data sources. Firstly, the National Alzheimer's Coordination Center (NACC) longitudinal dataset was used to train the ML model. Secondly, medical records gathered from Pauls Stradins Clinical University Hospital dating from 2020 to May 2025 were used to evaluate the algorithm's precision. Results: In the NACC study, the weighted four-class model achieved an overall accuracy of 84.0% and a balanced accuracy of 70.9%, but the SCD class remained poorly classified. After reframing the model to a three-class model the performance grew stronger for normal cognition, mild cognitive impairment (MCI) and dementia. Class distribution in the Latvian cohort consisted of dementia (n = 138); MCI (n = 13); and subjective cognitive decline (SCD) (n = 2). Dementia was identified most strongly-124/138 (sensitivity-89.9%). MCI was correct in 9/13 cases (sensitivity-69.2%). SCD cases were excluded. Overall, the model agreed with the neurologist-assigned diagnoses in 88.1% of the cases (133/151). Conclusions: The ML classification model has high precision when comparing with neurologist-assigned diagnoses, but it struggles to separate adjacent early-stage diagnoses, meaning that it did not reliably identify SCD. These findings support further methodological development and the implementation of prospective research. Nevertheless, this technology has high potential for being integrated in the future to aid triage and early screening, especially when advanced diagnostics are limited.
References: Geriatrics (Basel). 2022 Mar 08;7(2):. (PMID: 35314602)
Ageing Res Rev. 2024 Nov;101:102497. (PMID: 39293530)
Alzheimers Dement. 2014 Nov;10(6):844-52. (PMID: 24798886)
Alzheimers Dement. 2018 Jun;14(6):734-742. (PMID: 29352855)
J Alzheimers Dis. 2023;93(1):125-140. (PMID: 36938735)
Eur J Neurol. 2020 May;27(5):894-899. (PMID: 32043740)
Transl Psychiatry. 2021 Nov 9;11(1):576. (PMID: 34753917)
Int J Nanomedicine. 2019 Jul 19;14:5541-5554. (PMID: 31410002)
Cureus. 2023 Oct 4;15(10):e46454. (PMID: 37927664)
Sci Rep. 2022 Apr 8;12(1):5979. (PMID: 35395867)
NPJ Digit Med. 2022 Nov 7;5(1):171. (PMID: 36344814)
Alzheimers Res Ther. 2020 Dec 18;12(1):167. (PMID: 33339532)
Grant Information: P30 AG066515 United States AG NIA NIH HHS; P30 AG062421 United States AG NIA NIH HHS; P30 AG066508 United States AG NIA NIH HHS; P30 AG066519 United States AG NIA NIH HHS; P30 AG072973 United States AG NIA NIH HHS; P30 AG066462 United States AG NIA NIH HHS; P30 AG066530 United States AG NIA NIH HHS; P30 AG066509 United States AG NIA NIH HHS; P30 AG066546 United States AG NIA NIH HHS; P30 AG072958 United States AG NIA NIH HHS; P30 AG072972 United States AG NIA NIH HHS; P30 AG072979 United States AG NIA NIH HHS; P20 AG068082 United States AG NIA NIH HHS; P30 AG072975 United States AG NIA NIH HHS; P30 AG066444 United States AG NIA NIH HHS; P30 AG066507 United States AG NIA NIH HHS; P30 AG072946 United States AG NIA NIH HHS; P30 AG066518 United States AG NIA NIH HHS; P30 AG066511 United States AG NIA NIH HHS; P30 AG086404 United States AG NIA NIH HHS; P30 AG066512 United States AG NIA NIH HHS; P30 AG072978 United States AG NIA NIH HHS; P30 AG062429 United States AG NIA NIH HHS; P30 AG062422 United States AG NIA NIH HHS; R01 AG079280 United States AG NIA NIH HHS; P30 AG086401 United States AG NIA NIH HHS; P30 AG072977 United States AG NIA NIH HHS; P30 AG062677 United States AG NIA NIH HHS; P30 AG062715 United States AG NIA NIH HHS; P30 AG066506 United States AG NIA NIH HHS; P30 AG066468 United States AG NIA NIH HHS; P30 AG072976 United States AG NIA NIH HHS; P30 AG072947 United States AG NIA NIH HHS; P30 AG072931 United States AG NIA NIH HHS; P30 AG066514 United States AG NIA NIH HHS; P30 AG072959 United States AG NIA NIH HHS
Contributed Indexing: Keywords: NACC dataset; SHAP analysis; XGBoost classification; early dementia detection; machine learning algorithm; mild cognitive impairment; subjective cognitive decline
Entry Date(s): Date Created: 20260728 Date Completed: 20260729 Latest Revision: 20260824
Update Code: 20260824
PubMed Central ID: PMC13414067
DOI: 10.3390/medicina62071356
PMID: 42512898
Database: MEDLINE
FullText Text:
  Availability: 0
CustomLinks:
  – Url: https://www.doi.org/10.3390/medicina62071356?
    Name: ScienceDirect (all content) (s7799221)
    Category: fullText
    Text: View record from ScienceDirect
    MouseOverText: View record from ScienceDirect
  – Url: https://resolver.ebsco.com/c/fiv2js/result?sid=EBSCO:cmedm&genre=article&issn=16489144&ISBN=&volume=62&issue=7&date=20260714&spage=&pages=&title=Medicina (Kaunas, Lithuania)&atitle=Development%20and%20External%20Validation%20of%20a%20Machine%20Learning%20Model%20for%20Classification%20of%20Mild%20Cognitive%20Impairment%20and%20Dementia%20Using%20Clinical%20Data.&aulast=Kannenieks%20D&id=DOI:10.3390/medicina62071356
    Name: Full Text Finder (for New FTF UI) (ns324271)
    Category: fullText
    Text: Full Text Finder
    MouseOverText: Full Text Finder
Header DbId: cmedm
DbLabel: MEDLINE
An: 42512898
AccessLevel: 3
PubType: Academic Journal
PubTypeId: academicJournal
PreciseRelevancyScore: 0
IllustrationInfo
Items – Name: Title
  Label: Title
  Group: Ti
  Data: Development and External Validation of a Machine Learning Model for Classification of Mild Cognitive Impairment and Dementia Using Clinical Data.
– Name: Author
  Label: Authors
  Group: Au
  Data: <searchLink fieldCode="AU" term="%22Kannenieks+D%22">Kannenieks D</searchLink>; Faculty of Medicine, Riga Stradins University, LV-1007 Riga, Latvia.<br /><searchLink fieldCode="AU" term="%22Priede+Z%22">Priede Z</searchLink>; Department of Neurology and Neurosurgery, Riga Stradins University, LV-1002 Riga, Latvia.<br /><searchLink fieldCode="AU" term="%22Millers+A%22">Millers A</searchLink>; Department of Neurology and Neurosurgery, Riga Stradins University, LV-1002 Riga, Latvia.<br /><searchLink fieldCode="AU" term="%22Velins+KK%22">Velins KK</searchLink>; Department of Mathematics, Imperial College London, London SW7 2AZ, UK.
– Name: TitleSource
  Label: Source
  Group: Src
  Data: <searchLink fieldCode="JN" term="%229425208%22">Medicina (Kaunas, Lithuania)</searchLink> [Medicina (Kaunas)] 2026 Jul 14; Vol. 62 (7). <i>Date of Electronic Publication: </i>2026 Jul 14.
– Name: TypePub
  Label: Publication Type
  Group: TypPub
  Data: Journal Article
– Name: Language
  Label: Language
  Group: Lang
  Data: English
– Name: TitleSource
  Label: Journal Info
  Group: Src
  Data: <i>Publisher: </i><searchLink fieldCode="PB" term="%22MDPI%22">MDPI </searchLink><i>Country of Publication: </i>Switzerland <i>NLM ID: </i>9425208 <i>Publication Model: </i>Electronic <i>Cited Medium: </i>Internet <i>ISSN: </i>1648-9144 (Electronic) <i>Linking ISSN: </i><searchLink fieldCode="IS" term="%221010660X%22">1010660X </searchLink><i>NLM ISO Abbreviation: </i>Medicina (Kaunas) <i>Subsets: </i>MEDLINE
– Name: PublisherInfo
  Label: Imprint Name(s)
  Group: PubInfo
  Data: <i>Publication</i>: 2018- : Basel, Switzerland : MDPI<br /><i>Original Publication</i>: Kaunas : Lietuvos gydytojų sąjunga
– Name: SubjectMESH
  Label: MeSH Terms
  Group: Su
  Data: <searchLink fieldCode="MM" term="%22Cognitive+Dysfunction%22">Cognitive Dysfunction*</searchLink>/<searchLink fieldCode="MM" term="%22Cognitive+Dysfunction+classification%22">classification</searchLink> <br /><searchLink fieldCode="MM" term="%22Cognitive+Dysfunction%22">Cognitive Dysfunction*</searchLink>/<searchLink fieldCode="MM" term="%22Cognitive+Dysfunction+diagnosis%22">diagnosis</searchLink> <br /><searchLink fieldCode="MM" term="%22Machine+Learning%22">Machine Learning*</searchLink>/<searchLink fieldCode="MM" term="%22Machine+Learning+standards%22">standards</searchLink> <br /><searchLink fieldCode="MM" term="%22Dementia%22">Dementia*</searchLink>/<searchLink fieldCode="MM" term="%22Dementia+classification%22">classification</searchLink> <br /><searchLink fieldCode="MM" term="%22Dementia%22">Dementia*</searchLink>/<searchLink fieldCode="MM" term="%22Dementia+diagnosis%22">diagnosis</searchLink><br /><searchLink fieldCode="MH" term="%22Humans%22">Humans</searchLink> ; <searchLink fieldCode="MH" term="%22Aged%22">Aged</searchLink> ; <searchLink fieldCode="MH" term="%22Retrospective+Studies%22">Retrospective Studies</searchLink> ; <searchLink fieldCode="MH" term="%22Male%22">Male</searchLink> ; <searchLink fieldCode="MH" term="%22Female%22">Female</searchLink> ; <searchLink fieldCode="MH" term="%22Boosting+Machine+Learning+Algorithms%22">Boosting Machine Learning Algorithms</searchLink> ; <searchLink fieldCode="MH" term="%22Aged%2C+80+and+over%22">Aged, 80 and over</searchLink> ; <searchLink fieldCode="MH" term="%22Cohort+Studies%22">Cohort Studies</searchLink> ; <searchLink fieldCode="MH" term="%22Classification+Algorithms%22">Classification Algorithms</searchLink> ; <searchLink fieldCode="MH" term="%22Predictive+Learning+Models%22">Predictive Learning Models</searchLink>
– Name: Abstract
  Label: Abstract
  Group: Ab
  Data: Background and Objectives: As society ages, the number of patients with cognitive impairment is increasing. Machine learning methods that use structured clinical and cognitive-assessment data during routine diagnostic work-up may support and monitor structured classifications of cognitive status. This kind of approach can improve early screening, reduce physicians' workload and develop greater support for personalized treatment. To develop an XGBoost-based machine learning model using the National Alzheimer's Coordinating Center (NACC) dataset and to evaluate the model's precision with clinician-assigned diagnosis in a Latvian retrospective cohort study. Materials and Methods: The research was designed as a retrospective external validation cohort study that used two data sources. Firstly, the National Alzheimer's Coordination Center (NACC) longitudinal dataset was used to train the ML model. Secondly, medical records gathered from Pauls Stradins Clinical University Hospital dating from 2020 to May 2025 were used to evaluate the algorithm's precision. Results: In the NACC study, the weighted four-class model achieved an overall accuracy of 84.0% and a balanced accuracy of 70.9%, but the SCD class remained poorly classified. After reframing the model to a three-class model the performance grew stronger for normal cognition, mild cognitive impairment (MCI) and dementia. Class distribution in the Latvian cohort consisted of dementia (n = 138); MCI (n = 13); and subjective cognitive decline (SCD) (n = 2). Dementia was identified most strongly-124/138 (sensitivity-89.9%). MCI was correct in 9/13 cases (sensitivity-69.2%). SCD cases were excluded. Overall, the model agreed with the neurologist-assigned diagnoses in 88.1% of the cases (133/151). Conclusions: The ML classification model has high precision when comparing with neurologist-assigned diagnoses, but it struggles to separate adjacent early-stage diagnoses, meaning that it did not reliably identify SCD. These findings support further methodological development and the implementation of prospective research. Nevertheless, this technology has high potential for being integrated in the future to aid triage and early screening, especially when advanced diagnostics are limited.
– Name: Ref
  Label: References
  Group: RefInfo
  Data: Geriatrics (Basel). 2022 Mar 08;7(2):. (PMID: <searchLink fieldCode="PM" term="%2235314602%22">35314602)</searchLink><br />Ageing Res Rev. 2024 Nov;101:102497. (PMID: <searchLink fieldCode="PM" term="%2239293530%22">39293530)</searchLink><br />Alzheimers Dement. 2014 Nov;10(6):844-52. (PMID: <searchLink fieldCode="PM" term="%2224798886%22">24798886)</searchLink><br />Alzheimers Dement. 2018 Jun;14(6):734-742. (PMID: <searchLink fieldCode="PM" term="%2229352855%22">29352855)</searchLink><br />J Alzheimers Dis. 2023;93(1):125-140. (PMID: <searchLink fieldCode="PM" term="%2236938735%22">36938735)</searchLink><br />Eur J Neurol. 2020 May;27(5):894-899. (PMID: <searchLink fieldCode="PM" term="%2232043740%22">32043740)</searchLink><br />Transl Psychiatry. 2021 Nov 9;11(1):576. (PMID: <searchLink fieldCode="PM" term="%2234753917%22">34753917)</searchLink><br />Int J Nanomedicine. 2019 Jul 19;14:5541-5554. (PMID: <searchLink fieldCode="PM" term="%2231410002%22">31410002)</searchLink><br />Cureus. 2023 Oct 4;15(10):e46454. (PMID: <searchLink fieldCode="PM" term="%2237927664%22">37927664)</searchLink><br />Sci Rep. 2022 Apr 8;12(1):5979. (PMID: <searchLink fieldCode="PM" term="%2235395867%22">35395867)</searchLink><br />NPJ Digit Med. 2022 Nov 7;5(1):171. (PMID: <searchLink fieldCode="PM" term="%2236344814%22">36344814)</searchLink><br />Alzheimers Res Ther. 2020 Dec 18;12(1):167. (PMID: <searchLink fieldCode="PM" term="%2233339532%22">33339532)</searchLink>
– Name: GrantInfo
  Label: Grant Information
  Group: Grant
  Data: P30 AG066515 United States AG NIA NIH HHS; P30 AG062421 United States AG NIA NIH HHS; P30 AG066508 United States AG NIA NIH HHS; P30 AG066519 United States AG NIA NIH HHS; P30 AG072973 United States AG NIA NIH HHS; P30 AG066462 United States AG NIA NIH HHS; P30 AG066530 United States AG NIA NIH HHS; P30 AG066509 United States AG NIA NIH HHS; P30 AG066546 United States AG NIA NIH HHS; P30 AG072958 United States AG NIA NIH HHS; P30 AG072972 United States AG NIA NIH HHS; P30 AG072979 United States AG NIA NIH HHS; P20 AG068082 United States AG NIA NIH HHS; P30 AG072975 United States AG NIA NIH HHS; P30 AG066444 United States AG NIA NIH HHS; P30 AG066507 United States AG NIA NIH HHS; P30 AG072946 United States AG NIA NIH HHS; P30 AG066518 United States AG NIA NIH HHS; P30 AG066511 United States AG NIA NIH HHS; P30 AG086404 United States AG NIA NIH HHS; P30 AG066512 United States AG NIA NIH HHS; P30 AG072978 United States AG NIA NIH HHS; P30 AG062429 United States AG NIA NIH HHS; P30 AG062422 United States AG NIA NIH HHS; R01 AG079280 United States AG NIA NIH HHS; P30 AG086401 United States AG NIA NIH HHS; P30 AG072977 United States AG NIA NIH HHS; P30 AG062677 United States AG NIA NIH HHS; P30 AG062715 United States AG NIA NIH HHS; P30 AG066506 United States AG NIA NIH HHS; P30 AG066468 United States AG NIA NIH HHS; P30 AG072976 United States AG NIA NIH HHS; P30 AG072947 United States AG NIA NIH HHS; P30 AG072931 United States AG NIA NIH HHS; P30 AG066514 United States AG NIA NIH HHS; P30 AG072959 United States AG NIA NIH HHS
– Name: SubjectMinor
  Label: Contributed Indexing
  Group:
  Data: <i>Keywords: </i>NACC dataset; SHAP analysis; XGBoost classification; early dementia detection; machine learning algorithm; mild cognitive impairment; subjective cognitive decline
– Name: DateEntry
  Label: Entry Date(s)
  Group: Date
  Data: <i>Date Created: </i>20260728 <i>Date Completed: </i>20260729 <i>Latest Revision: </i>20260824
– Name: DateUpdate
  Label: Update Code
  Group: Date
  Data: 20260824
– Name: PubmedCentralID
  Label: PubMed Central ID
  Group: ID
  Data: PMC13414067
– Name: DOI
  Label: DOI
  Group: ID
  Data: 10.3390/medicina62071356
– Name: AN
  Label: PMID
  Group: ID
  Data: 42512898
PLink https://search.ebscohost.com/login.aspx?direct=true&site=eds-live&db=cmedm&AN=42512898
RecordInfo BibRecord:
  BibEntity:
    Identifiers:
      – Type: doi
        Value: 10.3390/medicina62071356
    Languages:
      – Code: eng
        Text: English
    Subjects:
      – SubjectFull: Humans
        Type: general
      – SubjectFull: Aged
        Type: general
      – SubjectFull: Retrospective Studies
        Type: general
      – SubjectFull: Male
        Type: general
      – SubjectFull: Female
        Type: general
      – SubjectFull: Boosting Machine Learning Algorithms
        Type: general
      – SubjectFull: Aged, 80 and over
        Type: general
      – SubjectFull: Cohort Studies
        Type: general
      – SubjectFull: Classification Algorithms
        Type: general
      – SubjectFull: Predictive Learning Models
        Type: general
      – SubjectFull: Cognitive Dysfunction classification
        Type: general
      – SubjectFull: Cognitive Dysfunction diagnosis
        Type: general
      – SubjectFull: Machine Learning standards
        Type: general
      – SubjectFull: Dementia classification
        Type: general
      – SubjectFull: Dementia diagnosis
        Type: general
    Titles:
      – TitleFull: Development and External Validation of a Machine Learning Model for Classification of Mild Cognitive Impairment and Dementia Using Clinical Data.
        Type: main
  BibRelationships:
    HasContributorRelationships:
      – PersonEntity:
          Name:
            NameFull: Kannenieks D
      – PersonEntity:
          Name:
            NameFull: Priede Z
      – PersonEntity:
          Name:
            NameFull: Millers A
      – PersonEntity:
          Name:
            NameFull: Velins KK
    IsPartOfRelationships:
      – BibEntity:
          Dates:
            – D: 14
              M: 07
              Text: 2026 Jul 14
              Type: published
              Y: 2026
          Identifiers:
            – Type: issn-electronic
              Value: 1648-9144
          Numbering:
            – Type: volume
              Value: 62
            – Type: issue
              Value: 7
          Titles:
            – TitleFull: Medicina (Kaunas, Lithuania)
              Type: main
ResultId 1