| Objective:Late-onset depression(LOD)and Mild cognitive impairment(MCI)patients,who are at high risk for Alzheimer’s disease(AD),have become closely associated with each other.The pattern of cognitive impairment and its possible pathophysiological mechanisms are still unclear and further studies are necessary.Therefore,this study used a case-control study to explore the differences between LOD and MCI patients in different cognitive domains based on a multidimensional neuropsychological test,and used Functional near-infrared spectroscopy(fNIRS)to find out whether there are corresponding objective markers of LOD and MCI.The study also aims to identify the possible specific cognitive differences between patients with LOD and MCI,and to clarify the cognitive impairment patterns and pathophysiological mechanisms of both,so as to provide a scientific basis for early clinical identification and intervention in the cognitive impairment population.Methods:A total of 46 patients with LOD,13 patients with MCI and 14 normal controls(HC)were included in this study.The study subjects who met the inclusion criteria were clinically diagnosed by at least two attending psychiatrists or above according to the Diagnostic and Statistical Manual of Mental Disorders(DSM-5)and the 2004 Petersen Diagnostic Criteria for Mild Cognitive Impairment.General demographic data,clinical assessment scales,neuropsychological tests and fNIRS data were collected by trained researchers.All neuropsychological tests were integrated into six cognitive domains,including memory,information processing speed,visuospatial function,executive function,language function,and social cognitive function.To reduce random variability to increase statistical power,raw scores of the different cognitive domain tests were converted into standardised Z-scores in this study.The combined Z-score for each cognitive domain was the average of the Z-scores of all subtests in that domain.The fNIRS spectral signals were pre-processed using MATLAB R2013 b software fNIRS＿SPM toolbox to extract the β values for channels ch1-ch52.The general demographic information,clinical assessment scales,neuropsychological test scores,andβ values of the fNIRS ch1-ch52 channels were compared between the three study groups using SPSS 25.0.If the differences between the three groups were statistically significant,Bonferroni post hoc tests were used to analyse the differences between the two groups,and chi-square tests were used to compare the differences between the three groups for categorical variables(e.g.gender)between the three groups and differences were considered statistically significant when P < 0.05.In addition,the β values extracted from the fNIRS channels that differed in the LOD and MCI groups were analysed in correlation with the Clinical Assessment Scale scores(HAMD-24,HAMA,MMSE,Mo CA)and the Z values of different cognitive domains,respectively,and the Z values of different cognitive domains in the LOD group were analysed in correlation with the HAMD-24 scores to explore the correlation between beta values of abnormal channels and clinical symptoms and cognitive function in LOD and MCI patients,and the relationship between cognitive function and severity of depressive symptoms in LOD patients.Results:1.Between-group comparisons of general demographic information and clinical symptoms: the differences between the three groups in terms of age(P = 0.160),gender(P = 0.328)and years of education(P = 0.135)were not statistically significant and were balanced and comparable.In terms of clinical symptoms,the LOD group had significantly higher HAMD-24 scores(LOD vs.MCI: P < 0.001;LOD vs.HC: P < 0.001)and HAMA scores(LOD vs.MCI: P < 0.001;LOD vs.HC: P < 0.001)than the MCI and HC groups,while the HAMD-24 scores(MCI vs.HC: P = 1.000),and HAMA scores(MCI vs.HC: P = 1.000)between MCI and HC scores did not differ statistically significantly.In terms of overall cognitive function,the differences between the three groups were statistically significant in terms of total MMSE score(P < 0.001)and total Mo CA score(P < 0.001),with the LOD and MCI groups having a statistically significant difference compared to the HC group in terms of total MMSE score(LOD vs.HC: P <0.001;MCI vs.HC: P = 0.035)and total Mo CA score(LOD vs.HC: P < 0.001;MCI vs.HC: P = 0.012)were significantly lower,while the differences in total MMSE scores(LOD vs.MCI: P = 0.931)and Mo CA scores(LOD vs.MCI: P = 0.206)between the LOD and MCI were not statistically significant.2.Between-group comparison of neuropsychological tests and cognitive functions:the three study groups differed in six cognitive domains,respectively,in memory(P <0.001),information processing speed(P = 0.005),visuospatial function(P < 0.001),executive function(P = 0.004),language function(P < 0.001)and social cognitive function(P = 0.002)with statistically significant.In particular,compared to the HC group,the LOD group showed a significant decrease in memory(P < 0.001),information processing speed(P = 0.003),visuospatial function(P < 0.001),executive function(P =0.003),language function(P < 0.001)and social cognitive function(P = 0.002);the MCI group showed a significant decrease in visuospatial function(P = 0.021)and language function(P = 0.023),while there was no statistically significant difference between the two groups in memory(P = 0.120),information processing speed(P = 0.127),executive function(P = 0.126)and social cognitive function(P = 0.093).In addition,patients in the LOD group had significantly lower memory(P = 0.003)than those in the MCI group,and there were no statistically significant differences between the two groups in information processing speed(P = 1.000),visuospatial function(P = 1.000),executive function(P = 1.000),language function(P = 0.186),or social cognitive function(P =1.000).3.Results of fNIRS analysis: The results showed that the extracted beta values in11 channels distributed in the prefrontal-temporal cortex(ch3,ch5,ch7-8,ch10,ch19,ch21,ch36,ch46-48)differed statistically between the three groups(ch3: P = 0.041;ch5:P = 0.008;ch7: P = 0.01;ch8: P = 0.041;ch10: P = 0.004;ch19: P = 0.017;ch21: P =0.002;ch36: P = 0.038;ch46: P = 0.039;ch47: P = 0.045;ch48: P = 0.002),while in the remaining lanes the differences were not statistically significant.Further comparison of the β values of the above channels between groups by post hoc analysis showed that in channels ch3(P = 0.043),ch5(P = 0.037),ch7(P = 0.018),ch8(P = 0.046),ch10(P =0.004),ch19(P = 0.014),ch21(P = 0.003),the LOD group had significantly lower βvalues were significantly lower than those in the HC group,and in the ch36(P = 0.036),ch48(P = 0.010)channels the LOD group had significantly lower β values than the MCI group.In addition,the differences in β values between the MCI and HC groups in the ch21(P = 0.010),ch46(P = 0.038),ch47(P = 0.043)channels were statistically significant.4.Correlation of β values of each differential channel of fNIRS with clinical symptoms and cognitive function: β values of each differential channel of the LOD group were correlated with the HAMD-24,HAMA,MMSE,Mo CA and other clinical symptom assessment scales and cognitive function,respectively.ch36 β values of the LOD group were positively correlated with social cognitive function(r = 0.315,P = 0.033);ch10 βvalues were positively correlated with information processing speed(r = 0.351,P =0.017),executive function(r = 0.317,P = 0.032),and social cognitive function(r =0.331,P = 0.025);ch5 β values were positively correlated with social cognitive function(r = 0.338,P = 0.022),MMSE scores(r = 0.294,P = 0.048),and Mo CA scores(r =0.322,P = 0.029);ch19 β values in the LOD group were negatively correlated with visuospatial function(r =-0.306,P = 0.039);and ch8 β values were negatively correlated with memory(r =-0.295,P = 0.046)and visuospatial function(r =-0.356,P = 0.015).In addition,correlation analysis of HAMD-24 scores with overall cognitive function and cognitive function across cognitive domains in the LOD group showed that HAMD-24 scores were negatively correlated with total MMSE score(r =-0.324,P = 0.028),total MOCA score(r =-0.407,P = 0.005),and visuospatial function(r =-0.342,P = 0.020).In the MCI group,β values for each differential channel were correlated with clinical symptom scores and cognitive function,respectively.The MCI group showed a negative correlation between ch48 β values and visuospatial function(r =-0.622,P = 0.023),with no significant correlation seen for the rest.Conclusion:This study found differences in cognitive impairment between LOD and MCI patients,as well as varying levels of metabolic brain activity.Overall,LOD showed a wider range of cognitive impairment domains than MCI patients,with LOD showing cognitive impairment in six different cognitive domains: memory,information processing speed,visuospatial function,executive function,language function,and social cognitive function,with memory impairment being more significant,and overall cognitive function and visuospatial function being negatively correlated with the severity of depressive symptoms in LOD patients.Both the MCI and LOD groups showed significant decreases in visuospatial function and language function,which may be a common feature of cognitive impairment in both groups.In addition,fNIRS imaging showed that metabolic activation impairment in the frontotemporal cortex may be more extensive in LOD patients than in MCI patients and correlated with cognitive function,which was further validated on the basis of clinical assessment.The results of this study provide new insights into the specificity of cognitive impairment in LOD and MCI,clarify the cognitive impairment patterns of both and their pathophysiological mechanisms,and provide some scientific basis for early clinical identification and intervention in cognitively impaired populations. |
PDF Full Text Request