BackgroundExecutive function deficits (EFD) in late‐life depression (LLD) has been reported to be associated with antidepressant treatment resistance, increased disability, and poor quality of life. However, the underlying neutral mechanisms of EFD in patients with the first episode of LLD remains unclear.MethodsA total of 27 patients with first‐episode, drug‐naive LLD and 27 non‐depressed controls (NC) were recruited for the present research. Participants underwent the Trail Making Test, the 17‐item Hamilton depression rating scale (HAMD‐17) test, and task‐state functional magnetic resonance imaging scans under the neutral Stroop task. LLD patients' executive functions, depressive symptoms, and brain activity were examined again after 6 months of antidepressant treatment.ResultsOf the 27 LLD patients, 16 cases completed 6‐month follow‐ups. Patients in the LLD baseline group spent more time on the Trail Making Test A test than those in the NC group (p < 0.05). In the presence of an incongruency between the word color and meaning, the accuracy rate of the neutral Stroop task in the LLD baseline group was lower, and the reaction time was greater than that in the NC group, with statistically significant difference (p < 0.05). The HAMD‐17 score in the LLD follow‐up group was significantly lower than that in the LLD baseline group (p < 0.05). More activated brain regions were present in the LLD baseline group than in the NC group when performing the neutral Stroop task. Compared with the LLD baseline group, abnormal activation of relevant brains in the cingulate‐prefrontal‐parietal network of LLD patients still existed in the LLD follow‐up group.ConclusionsLLD patients engaged more brain areas than the NC group while performing the neutral Stroop task. Abnormal activation of the cingulate‐prefrontal‐parietal network could be a contributing factor to EFD in LLD.Trial registrationChiCTR, ChiCTR2100042370 (Date of registration: 21/01/2021).LimitsWe didn't enroll enough first‐episode, LLD patients, the robustness of the findings need to be confirmed by large sample clinical trials.