TABLE 1.

Summary of the studies assessing fNIRS measures during cognitive tasks in older adults.

First author (Journal, year), Country	Sample size (N) Clinical characteristics of the sample (mean age +SD) NIRS optodes localization	Paradigm description	Main fNIRS results
Verbal fluency
Heinzel (Neurobiol Aging, 2013), Germany (Heinzel et al., 2013)	N = 325 Healthy (64.6 ± 7.3). Prefrontal, temporal and parietal.	Tasks: – Phonetic verbal fluency. –Semantic verbal fluency. –Control task: reciting week days. Three trials: 30 s each. Rest: 30 sec after each trial.	With increasing age: ⇓: Lower activation on bilateral inferior frontal junction during verbal fluency. Increased bilateral activation at middle frontal gyri and supramarginal gyri.
Heinzel (PLoS One 2015), Germany (Heinzel et al., 2015)	N = 1052 Healthy (65.2 ± 6.8). Prefrontal, parietal and fronto-temporal.	Tasks: –Phonetic verbal fluency. –Semantic verbal fluency. –Control task: reciting week days. Three trials: 30 s each. Rest: 30 sec after each trial.	⇑: Increased activation during both verbal fluencies (compared to control task). Stronger response in phonological than semantic (increased activation in right prefrontal and bilateral inferior parietal regions extending toward postcentral gyri and decreased in bilateral fronto-temporal areas).
Yeung (Front Aging Neurosci., 2016), China (Yeung et al., 2016a)	N = 52 MCI (69.1 ± 6.2); Healthy (68.8 ± 6.1). Prefrontal.	Tasks: – Semantic verbal fluency. Two task blocks, 60 s each. Rest: repeat “1, 2, 3, 4” out loud. Before and after VF task.	⇑: Increased O2Hb bilaterally during verbal fluency in both groups. No significant group differences. Control group showed left lateralization of frontal lobe activation (whereas MCI group did not).
Yap (Front Aging Neurosci., 2017), Malaysia (Yap et al., 2017)	N = 61 Healthy (72.6 ± 8.5); MCI (73.1 ± 8.2); Mild AD (74.7 ± 10). Prefrontal and part of temporal.	Task: – Semantic verbal fluency (60 s). Rest: 20 s before and after task.	⇑: Highest O2Hb increase during task was observed in MCI followed by healthy and mild AD.
Katzorke (Psychiatry Res Neuroimaging, 2018), Germany (Katzorke et al., 2018)	N = 110 Healthy (74.2 ± 1.6); MCI (74.0 ± 1.6). Fronto-temporal.	Tasks: – Letter verbal fluency. – Semantic verbal fluency. – Control (weekdays). Three trials per task: 30 sec each. Rest: 30 s after each task.	⇓: Decreased PFC activation during semantic verbal fluency in MCI compared to healthy controls (but not during phonological verbal fluency).
N-back tasks
Niu (CNS Neurosci Ther., 2013), China (Niu et al., 2013)	N = 24 MCI (64.8 ± 7.2); healthy (63.5 ± 5.3). Frontal, parietal and temporal.	Task: –Digit N-back task (0-back and 1-back conditions). Three blocks in each condition: 20 trials for each block followed by 1000 ms interstimulus period. Rest: No rest time specified.	⇓: MCI participants showed lower O2Hb concentrations in the left dorsolateral PFC, right supplementary motor area and left superior temporal regions compared to control group.
Vermeij (Front Aging Neurosci., 2014), Netherlands (Vermeij et al., 2014)	N = 18 Healthy older adults (70.8 ± 5.0). Prefrontal.	Task: –Spatial N-back (0-back, 1-back and 2-back conditions). 60 trials (500 ms each) with 3000 ms interval between trials. Rest: Initial 1 min-baseline (staring at screen).	⇑: Increased working-memory load associated increased prefrontal activation and decreased performance.
Yeung (Dement Geriatr Cogn Disord., 2016), China (Yeung et al., 2016b)	N = 52 MCI (69.1 ± 6.3); cognitively normal (68.8 ± 6.1). Frontal.	Task: –Digit N-back task (0-back and 2-back conditions). 20 trials (1000 ms each) followed by 1000 ms interval between trials. Rest: 30 s between blocks.	⇓: MCI group did not show frontal activation. Tended to reduce activation with high working memory load. ⇑: Control group: frontal activation in high working memory load (2-back condition).
Other tests for executive functions
Heilbronner (Neuroimage, 2013), Germany (Heilbronner and Münte, 2013)	N = 35. Healthy older adults (68 ± 1.4); younger adults group (23.1 ± 0.4). Frontotemporal.	Task: –Cognitive Go/No Go inhibition task: Go stimulus: press button; No Go stimulus: inhibit pressing button. 1083 stimuli in 5 trials. Rest: Rest in a self-paced manner.	⇑: Older adults showed activation in frontal areas. Compared to young participants, activation shifted rostrally (left hemisphere) and dorsally (right hemisphere) in older adults.
∗ Albinet (Front Aging Neurosci., 2014), France (Albinet et al., 2014)	N = 40 Healthy old adults: high-fit (67.32 ± 4.48); low-fit (68.88 ± 3.87) Prefrontal	Task: –Random Number Generation: participants asked to say random number when heard a tone. Fast pace (tone/1 s) and slow pace (tone/1.5 s). Two trials of 100 responses at each pace. –Control: count in order from one to nine.	⇑: Increasing activation in relation to task difficulty. High-fit group showed greater increase in O2Hb.
Müller (Neuropsychologia, 2014), Germany (Müller et al., 2014)	N = 40 Older adults (70.9 ± 3.5); younger adults (25.7 ± 3.0). PFC, motor and premotor regions.	Task: –Adapted version of Trail Making Test (TMT) A and B. –Control task: retrace 90 interconnected circles. Each test presented three times (30 s each). Rest: 30 s after each test.	⇑: Older adults showed bilateral ventrolateral and dorsolateral prefrontal and premotor cortex activation during TMT-B (more channels active in the right hemisphere). Additional activation in medial and lateral PFC in elderly (younger participants show more ventral PFC, especially in the left hemisphere).
Oboshi (PLoS One, 2014), Japan (Oboshi et al., 2014)	N = 120 Healthy older adults (71.0 ± 6.4); younger adults (21.7 ± 3.3). Prefrontal.	Task: –Visual working memory task. Six blocks (28.8 s each). Rest: 30 s.	⇑: Older adults: O2Hb increase during working memory task. Young adults: Higher O2Hb increase during pre-task (compared to elderly). Both groups: lower activation during pre-task is associated with higher O2Hb change during working memory task.
∗ Dupuy (Front Hum Neurosci., 2015), Canada (Dupuy et al., 2015)	N = 58 Healthy older adults (62.9 ± 5.4); young adults (24.6 ± 3.6). Prefrontal.	Task: –Modified Stroop-task with two conditions: naming (identify the color of the ink); executive or incongruent (color of the ink not matching the color-word displayed). Four trial-blocks (60 s each). Rest: 60 s between blocks.	⇑: High-fit women showed increased activation in right inferior frontal gyrus (independent of age group).
∗ Hyodo (Neuroimage, 2015), Japan (Hyodo et al., 2016)	N = 60 Healthy older adults (70.3 ± 3.2). Prefrontal.	Task: –Modified Stroop-task: participants asked to decide if word is printed in the color written below the word (neutral and incongruent conditions). 60 trials (30 neutral and 30 incongruent trials). Rest: 9 –13 s interstimulus interval.	Higher fitness levels and left-lateralized PFC activation related to shorter Stroop interference time. Higher fitness associated with more left-lateralized activation.
Laguë-Beauvais (Brain and Cognition, 2015), Canada (Laguë-Beauvais et al., 2015)	N = 35 Healthy older adults (63.47 ± 3.67); young adults (23.94 ± 2.32). Prefrontal.	Task: –Color task: identify color of an “X” on screen (by typing on keyboard). –Letter task: identify “K” or “L” on screen (by typing on keyboard). ∙ Conditions: –Single pure: only an “X” or letter is displayed to perform one of the tasks. –Dual mixed: both an “X” and a letter are displayed and have to be answered. These are performed under the instruction to prioritize the letter over the color task (Priority Block) or to give the same priority to both tasks (Equal Block). Rest: staring at fixation cross on screen (1000 ms).	Priority condition: –Older adults: activation in the left dorsolateral prefrontal cortex and bilateral ventrolateral cortex during DT. –Young adults: dual mixed trials showed greater changes in more frontal areas, especially right sided. Equal condition: –Older adults: dual mixed trials engaged bilateral dorsolateral prefrontal cortex, compared to single trials. –Young adults: no differences between activation during dual mixed and single trials. Single trials showed change in activation in right posterior dorsolateral prefrontal cortex for HHb. The activation change between priority and equal conditions was found only in older adults.
Bierre (J Gerontol A Biol Sci Med Sci., 2017), New Zealand (Bierre et al., 2017)	N = 72 Healthy older adults (66 ± 3.8); young adults (21.9 ± 2.7). Frontal.	Task: –Visuomotor tasks (increasing executive demand): 1)Basic visuomotor performance. 2)Adding inhibition. 3)Adding need to switch between tasks. Rest: 2 min (sitting).	⇑: Older adults showed increased O2Hb in relation to increasing task difficulty. Older adults showed higher O2Hb compared to younger adults
Halliday (Neurophoton., 2017), Canada (Halliday et al., 2017)	N = 25 Healthy older adults (75.88 ± 3.28) Prefrontal.	Task: –Computerized cognitive task: Multi-Source Interference Task (congruent and incongruent condition). Fifteen trials in a 30 s block. Rest: 60 s baseline before task. 20 s between blocks.	Greater mean O2Hb during congruent (easier) task associated with faster performance and during incongruent (more difficult) task, with slower performance. Greater O2Hb variability at within-person level associated with better accuracy and faster performance. Greater O2Hb variability at between-person level associated with slower performance.
Huppert (PLoS One, 2017), USA (Huppert et al., 2017)	N = 19 Older adults (88.1 ± 6.0). Frontal.	Tasks: –Stroop Test. –Symbol Digit Coding. –Shifting Attention Test. Rest: 30 s (quiet sitting baseline).	⇑: Left Broadmann’s area (BA) 10 (right superior frontal) activation during Symbol Digit Coding and Shifting Attention Test. Right BA-10, right BA-45 and left BA-10 activated during Stroop test.
∗ Halliday (J Clin Exp Neuropsychol., 2018), Canada (Halliday et al., 2018)	N = 27 Older adults (76.1 ± 3.3). Prefrontal.	Task: –Computerized cognitive task: Multi-Source Interference Task (congruent and incongruent condition). Fifteen trials in a 30 s block (total of 4 blocks for each condition). Rest: No rest time specified.	⇑: Fallers: activation during congruent and incongruent task; recruited additional tissue to perform at similar level. Non-fallers: no active channels during congruent task; little activation during incongruent task (medial right prefrontal cortex).
Memory test
Uemura (Int J Geriatr Psychiatry, 2016), Japan (Uemura et al., 2016)	N = 130 Amnestic MCI (71.8 ± 43); healthy older adults (71.7 ± 3.9). Prefrontal.	Task: –Encoding and retrieval of 10 words (20–30 s respectively). Repeat vowels: Pre-task (10 s), rest after task (20–30 s) and post-task (10 s).	⇓: Reduced activation in bilateral dorsolateral cortex during memory retrieval in amnestic MCI. No significant group effects during encoding.

AD, Alzheimer’s Disease; DT, dual-task; fNIRS, Functional Near-Infrared Spectroscopy; HHb, deoxygenated hemoglobin; MCI, Mild Cognitive Impairment; O₂Hb, oxygenated hemoglobin; PFC, Prefrontal Cortex; ST, single task; SD, Standard Deviation; VF, Verbal Fluency. ^∗Articles assessing modulation of health characteristics on brain activation (Table 3).