Table 2 Overview of the characteristics of cognitive testing and the main outcomes of the reviewed studies

First author [ref.]

(1) Cognitive testing

(2) Main findings (related to functional and/or structural brain changes in response to resistance exercises or resistance training)

 Chang et al. [43]

(1) Executive functions (Stroop test) during fNIRS (conducted 15 min after exercise cessation)

(2) Between group comparisons (postexercise, neutral condition):

 - ↓ TOI in lt. PFC during CT (HIR vs. CON (n) / MIC)

 - ↑ Solved items and ↓ response time during CT (HIR vs. CON (n))

 Between group comparisons (postexercise, incongruent condition):

 - ↓ TOI in lt. PFC (HIR vs. CON (n) / MIC)

 - ↓ TOI in rt. PFC (HIR vs. CON (n) / MIC / HIA)

  (ROI: lt. and. rt. PFC)

 Coetsee et al. [44]

(1) Executive functions (Stroop test) during fNIRS

(2) Posttest vs. pretest:

 - ↓ OxyHb in lt. PFC in RT during CT (Stroop interference effect)

 - ↓ THI in lt. PFC in RT and MCT during CT (Stroop interference effect)

 - ↓ Reaction time in RT during CT (naming and executive condition)

  (ROI: lt. and rt. PFC)

 Hong et al. [188]

(1) Cognitive test battery (Stroop test, COWAT, DFDB; Rey 15-Item Memory Test) and resting EEG

(2) Posttest versus pretest:

 - ↓ Relative theta power (at F3) in MCI RT

 - ↑ Relative alpha power (at T3) in MCI RT

 - ↓ Relative theta power (at P3 and at P4) in HOA RT

 - DB scores were significantly higher in MCI RT than in MCI CON (at posttest)

 Özkaya et al. [194]

(1) Auditory task during EEG

(2) Posttest vs. pretest:

 - ↓ Latencies of N1 (at Fz) and N1 (at Cz) in RT and AT

 - ↑ Amplitudes of N1-P2, P2-N2 and N2-P3 (at Fz) and N1-P2 (at Cz) in RT

 Between group comparisons:

 - ↓ Absolute changes in latencies of P2 and N2 (at Fz and at Cz) in RT compared with AT and CON

 - ↑ Absolute changes in amplitudes of N1-P2, P2-N2, and N2-P3 (at Fz) and N1-P2 and N2-P3 (at Cz) in RT compared with AT and CON

 Tsai et al. [182]

(1) Executive functions (Go/No-Go task combined with the Eriksen Flanker paradigm) during EEG measurements (CT was conducted after exercise cessation when the participant’s body temperature and HR had returned to within 10% of pre-exercise levels, which was on average approximately 5 min after acute resistance exercise cessation.)

(2) Posttest vs. pretest:

 - ↑ P3 amplitude (i.e., at Fz, Cz, and Pz) in MIRT and HIRT during CT

 - ↓ Reaction time in MIRT and HIRT during CT (Go condition)

 - ↑ Accuracy in MIRT and HIRT during CT (incongruent No-Go condition)

 - ↑ Serum GH and serum IGF-1 in MIRE and HIRE (prior to cognitive testing at pretest vs. prior to cognitive testing at posttest)

 - ↓ Serum cortisol in MIRE and HIRE (prior to cognitive testing at pretest vs. prior cognitive testing at posttest)

 - ↓ Serum GH and serum IGF-1 in HIRE (prior to cognitive testing at posttest vs. after cognitive testing at posttest)

 - ↑ Serum GH in MIRE and HIRE, serum IGF in MIRE (prior to cognitive testing at pretest vs. after cognitive testing at posttest)

 - ↓ Serum cortisol in MIRE (prior to cognitive testing at pretest vs. after cognitive testing at posttest)

 - Lower serum cortisol levels were associated with higher P3 amplitude

 Tsai et al. [187]

(1) Executive functions (oddball task) during EEG measurements

(2) Between group comparisons:

 - ↑ P3a amplitude (i.e., at F3 and F4) and P3b amplitude (i.e., at Cz, Pz, and Oz) in RT during CT compared with CON (n)

 - ↑ Accuracy in RT during CT compared with CON (n)

 - ↓ Reaction time in RT during CT compared with CON (n)

 Posttest vs. pretest:

 - ↓ Reaction time in RT during CT

 - ↑ Serum IGF-1 levels in RT

 - ↓ Serum homocysteine levels in RT

 - Higher serum IGF-1 levels in RT were associated with the faster reaction times and larger P3b amplitudes

 Tsai et al. [195]

(1) Working memory (Memory span from WAIS-IV); executive functions (Flanker task) during EEG measurements (CT was conducted after exercise cessation when the participant’s body temperature and HR had returned to within 10% of pre-exercise levels, which was on average approximately 5 min after acute resistance exercise cessation.)

(2) Posttest vs. pretest:

 - ↑ P3 amplitudes (i.e., at Fz, Cz, and Pz, except the Pz electrode in RE) in AE and RE during CT (in all conditions)

 - ↓ Reaction time in AE and RT during CT (congruent and incongruent condition)

 - ↑ Serum IGF-1 in AE and RE; serum BDNF and serum VEGF in AE (prior to cognitive testing at pretest vs. prior to cognitive testing at posttest)

 - ↓ IGF-1 in AE and RE and serum BDNF in AE (prior to cognitive testing at posttest vs. after cognitive testing at posttest)

 - Lower P3 latency across all participants was associated with higher IGF-1 levels (prior to cognitive testing at posttest)

 Tsai et al. [191]

(1) Working memory (Memory span from WAIS-IV); executive functions (Task switching) during EEG measurements

(2) Posttest vs. pretest:

 - ↑ P3 amplitudes in AE and RT

 - ↓ Reaction time in AE and RT during CT (non-switching condition and switching condition)

 - ↑ Accuracy rate in AE and RT during CT (pure condition, non-switching condition, and switching condition)

 - ↑ Serum IGF-1 in RT and serum BDNF in AT

 - ↓ Serum TNF-α and serum IL-15 in RT and AT / ↑ serum TNF-α in CON

 - Higher levels of VO_2max are associated with higher levels of serum BDNF in RT and AT

 Vonk et al. [183]

(1) Executive functions (Stroop test) during EEG measurements (conducted 10 min, 20 min, 30 min, and 40 min after exercise cessation)

(2) Posttest vs. pretest:

 - ↓ Response time in RE and LM during CT (congruent and incongruent condition, 10 min after exercise cessation vs. pretest)

 - ↓ Response time in RE and LM during CT (congruent condition, 10 min vs. 30 min after exercise cessation)

 - ↓ Accuracy in RE and LM during CT (incongruent condition, 30 min after exercise cessation vs. pretest)

 - ↑ P3 amplitude in RE and LM during CT (incongruent condition, 30 min and 40 min after exercise cessation vs. pretest)

 - ↑ P3 amplitude in RE and LM during CT (congruent condition, 10 min and 30 min after exercise cessation vs. pretest)

 Yerokhin et al. [192]

(1) Cognitive test battery (Stroop test, FOME; CFT); executive functions (oddball paradigm) during EEG

(2) Posttest vs. pretest:

 - ↓ Beta asymmetry and ↓ N200 A asymmetry

 - ↑ Delta asymmetry

 - ↑ Figure delayed recall and Fuld immediate recall

 - Decreased N200 A asymmetry was significantly correlated with improvements in Fuld immediate and Fuld delayed recall

 - Increase in delta asymmetry was significantly correlated with an improvement in Fuld delayed recall

  (ROI: frontal lobe [FP1, FP2, F7, F8])

 Best et al. [184]

(1) Cognitive test battery (Stroop test, TMT A&B, DB, RAVLT, DSST)

(2) Between group comparisons:

 - ↓ Cortical WM atrophy 2x RT compared with BAT at 2-year follow-up

 - ↑ Executive functions in 1x RT compared with BAT considering changes from baseline to postintervention

 - ↑ Executive functions in 1x RT and 2x RT compared with BAT considering changes from baseline to a 2-year follow-up

 - ↑ Memory performance in 2x RT compared with BAT considering changes from baseline to 2-year follow-up

 - ↑ Peak muscle power in 2x RT compared with BAT considering changes from baseline to postintervention and to a 2-year follow-up

 Brinke et al. [197]

(1) Memory (RAVLT)

(2) Between group comparisons:

 - No significant differences between AT and RT in hippocampal volume after 26 weeks

 - ↑ Hippocampal volume in rt. and lt. hemisphere / total hippocampal volume in AT compared with AT after 26 weeks

 - Positive partial correlation between increase in left hippocampal volume and change in RAVLT (loss after interference condition)

 Bolandzadeh et al. [185]

(1) Executive functions (Stroop test)

(2) Between group comparisons:

 - ↓ Cortical WML volume 2x RT compared with BAT at 2-year follow-up

 - ↓ WML progression in 2x RT at postintervention was associated with maintenance of gait speed

 Kjølhede et al. [193]

(1) Working memory & auditory information processing speed (PASAT)

(2) Changes in cortical thickness in response to RT:

 - ↑ E.g., in subcentral sulcus and gyrus; anterior cingulate sulcus and gyrus, middle anterior cingulate sulcus and gyrus, inferior parietal angular gyrus, inferior temporal gyrus, middle temporal gyrus, temporal pole, superior circular sulcus of insula, superior and transverse occipital sulcus, inferior temporal sulcus, orbital H-shaped sulcus, inferior and superior parts of the precentral sulcus, inferior and superior temporal sulcus

 Between group comparisons regarding cortical thickness:

 - ↑ Anterior cingulate sulcus and gyrus, temporal pole, inferior temporal sulcus, orbital H-shaped sulcus in RT compared with WL after 24 weeks

 - Greater thickness in the temporal pole was correlated with lower EDSS scores

 Liu-Ambrose et al. [186]

(1) Cognitive test battery (Stroop test, TMT A&B, DFDB)

(2) Between group comparisons:

 - ↑ Stroop test performance in 1x RT and 2x RT compared with BAT at 2-year follow-up

 - ↑ Peak muscle power in 2x RT compared with BAT at postintervention and to a 2-year follow-up

 - ↓ Whole brain volume (from baseline) in 1x RT and 2x RT compared with BAT at a 2-year follow-up

 - Improvement in Stroop test performance during intervention was significantly associated with increased gait speed

 Liu-Ambrose et al. [45]

(1) Executive functions test (modified Eriksen Flanker task) during fMRI

(2) Between group comparisons:

 - ↑ Activation of the left anterior insula extending into the lateral orbital frontal cortex in 2x RT compared with BAT at posttest in the incongruent condition

 - ↓ Activation of the left anterior insula extending into the lateral orbital frontal cortex and anterior portion of the left middle temporal gyrus in 2x RT compared with BAT at posttest in the congruent condition

 - ↓ Reduction in interference score (better performance) in 2x RT compared with BAT

 Nagamatsu et al. [189]

(1) Cognitive test battery (Stroop test, TMT A&B, DFDB; EPT) and associative memory (memorizing face-scene pairs) during fMRI

(2) Between group comparisons:

 - ↑ Stroop test performance and associate memory task performance in RT compared with BAT at postintervention

 - ↑ Activation of the right lingual and occipital-fusiform gyri and the right frontal pole in 2x RT during CT compared with BAT at postintervention (encoding and recall of associations)

 - Higher hemodynamic activity in the right lingual gyrus was correlated with better performance in the associative memory test

 Suo et al. [190]

(1) Cognitive test battery (e.g. ADAS, TMT A&B, BVRT, COWAT, Category Fluency, SDMT, Logical Memory WMS-III, Matrices WMS-III, Similarities WMS-III)

(2) Between group comparisons:

 - ↓ ADAS-Cog score (i.e., improved cognition) at posttest in the RT groups compared with all other groups

 - ↑ Posterior cingulate cortex grey matter thickness at postintervention in RT groups compared with all other groups

 - ↓ White matter hyperintensities volumes in the rt. periventricular zone and the rt. parietal zone in RT groups compared with all other groups (significant when analyzed at the regional level / not-significant when whole brain-corrected)

 - Greater posterior cingulate cortex grey matter thickness was significantly correlated with lower ADAS-Cog score (i.e. improved cognition)

 Functional connectivity changes:

 - ↓ PC_FC connectivity with the left inferior temporal lobe and the anterior cingulate cortex in RT + SHAM / ↓ PC_FC connectivity between the PC and the anterior cingulate cortex in CCT + SHAM

 - ↓ PC_FC between the PC and the anterior cingulate cortex in RT + CCT

 - ↑ HIP_FC connectivity with the right middle frontal lobe and ↓ connectivity with the right inferior temporal lobe in RT + SHAM

 - ↑ HIP_FC connectivity between the hippocampus and the left superior frontal lobe in CCT + SHAM

 - ↑ Hippocampal–anterior cingulate cortex connectivity and the hippocampal–right superior frontal lobe connectivity in RT + CCT

 - ↑ Superior functional connectivity between the hippocampus and the superior frontal lobe is associated with improved memory domain performance