The Mechanism Of Regulating Amygdala Activity To Improve Insomnia Disorder Based On Real-time FMRI Neurofeedback

Background and purposeInsomnia disorder(ID)is a major public health problem affecting people’s health and quality of life.It often accompanied with mood disord-ers.The neurofeedback technique based on real-time functional magnetic re-sonance imaging(rtfMRI)is a novel and noninvasive methods,it maybe used to improve congnitive function as a new approach.Previous studies have confirmed that amygdala,as the key brain area of emotion production,ex-pression and negative emotion perception,may participate in the pathogenesis of ID.In this study,real-time functional magnetic resonance Imaging neurofeedback(rtfMRI-NF)was applied,which took the left amygdala as the target area,used the analysis of brain functional network topology attribute and resting state brain function to explore the neural mechanism and short term transfer effecting of rtfMRI-NF in regulating the amygdala acti-vity of ID patients to improve insomnia,so as to provide a new imaging basis for the clinical treatment of ID,and was expected to provide a new and non-invasive model of neurofeedback treatment of ID.Materials and methodsFourteen ID patients,who met the enrollment criteria,were recruited.They need rtfMRI-NF training once a week for three weeks.During the training,the left amygdala of ID patients was used as the brain target area,and the amygdala activity was regulated by memory happy autobiographical memories(AMS),so as to improve the green temperature bar representing amygdala hemodynamic signals from the basic level to the target level.All patients underwent Pittsburgh Sleep Quality Index(PSQI),Hamilton Depression Scale(HAMD),Hamilton Anxiety Scale(HAMA),20-item Toronto Alexithymia Scale(tas-20),15-item Emotional Contagion Scale(ECS),Positive Affect and Negative AffectScale,PANAS)before and after rtfmri-nf training.The six related clinical scales were used to evaluate the changes of insomnia,depression,anxiety,cognitive function and emotional experience ability in ID patients before and after training.According to the decrease of the scores of the main emotion related scale(HAMA score decreased by 50%)of ID patients after training,14 subjects were divided into two groups:the effective group(n=6,HAMA score decreased by≥50%)and the ineffective group(n=8,HAMA score decreased by less than 50%).Spm12,DPABI and GRETNA software were used to process fMRI data of resting state and task state before and after rtfmri-nf training.Results1 General materials and clinicalscals results of two groupsThere was no significant difference in age(t=2.1 7,P=0.051),gender(t=0.50,P=0.779),education years(t=-0.27,P=0.803)between the two groups(P>0.05).The positive emotion score of PAN AS in the effective group was significantly lower than that in the ineffective group(t=-2.735,P=0.018).After 3 sessions of rtfMRI-NF training,the interview results showed that the effective group had better compliance than the ineffective group in the feedback training.During the feedback process,patients in the effective group ID strictly followed the guidance and learned the AMs strategy of recalling happy memories,and independently regulated their emotions according to the visual feedback results.Moreover,when the feedback effect was not satisfactory,the patients continued to autonomously regulated their emotions according to the learned emotion regulation strategy.The ineffective group could not concentrate easily in the feedback process,and sometimes could not continue to carry out independent emotional regulation according to the learning strategy when the feedback effect was not satisfactory.Compared the relevant clinical scale scores which acquired after session 3 with the scale scores of baseline,the results showed that PSQI(t=-7.225,P=0.001),HAMA(t=-3.500,P=0.002)and HADA(t=-5.824,P=0.002)scores in the effective group after the 3rd rtfMRI-NF training(session 3)showed significant statistical differences from the baseline period,while SHAPS,PANAS positive,PANAS negative,TAS,ESC showed no significant statistical differences from the baseline period.PSQI(t=-3.674,P=0.008)and HAMA(t=-2.478,P=0.042)score after the 3rd rtfMRI-NF training(session 3),in the invalid group showed significant statistical differences from the baseline period,and HADA score showed a downward trend from the baseline period,while PSQI,SHAPS,PANAS positive,PANAS negative,TAS,ESC showed no significant statistical differences from the baseline period.Compared the relevant clinical scale scores which acquired after transfer training with the scale scores of session 3,the results showed that HAMD(t=-3.464,P=0.018)scores in the effective group after the transfer training(session 4)showed significant statistical differences from the baseline period 3rd rtfMRI-NF training(session 3).HAMA score showed a downward trend compared with the end of session 3.However,there were no significant statistical differences in SHAPS,PANAS positive and PANAS negative,TAS and ESC.There was no significant statistical difference in the scores of 6 related scales after the end of session 4 in the invalid group compared with that after the end of session 3.2 Functional magnetic resonance results of two groups2.1 ROI Analysis of task state fMRI dataThrough ROI analysis,it was verified that the activity of bilateral amygdala in the effective group was different from that in the ineffective group.In the effective group,the activity of left amygdala increased in run1,run2 and run3 of three sessions,and the difference was statistically significant(sessionl:t(6))=3.4,P=0.001;session2:t(6))=2.8,P=0.009;session3:t(6))=1.7,P=0.046].The activity of right amygdala increased in run1,run2 and run3 of three sessions,but there was no significant difference between run3 and run1[session1:t(8)=2.6,P=0.895;session2:t(8)=1.6,P=0.967;session3:t(8)=-5.4,P=0.006].Considering that patients with ID may need adaptive feedback training in Run1 during rtfmri-nf training,and may be unable to concentrate due to fatigue in RuN3,we focused on the analysis of the average β weight of run2 in the left amygdala of the effective group and the ineffective group.The results showed that the activity of run2 in the three sessions of the effective group increased.The activation of the left amygdala in the third rtfmri-nf training(Session 3)run2 and the first rtfmri-nf training(Session 2)run2 of the effective group and the ineffective group were analyzed by post analysis.It was found that the activity of the left amygdala in the effective group was significantly higher than that in the run2 of the session 1[t(6)=1.2,P=0.042),but the activity of the left amygdala of three sessions in the invalid group increased,but the activity did not increase significantly.There was a statistical difference between the activity of the left amygdala of session 3 and that of session 1[t(8)=-4.2,P=0.035)](see the attached figure,figure 5 for details).These results showed that the activity of amygdala was significantly increased in the effective group,but not in the ineffective group.In order to observe the transmission effect of rtmri-nf training on amygdala activity,the β weight of transfer in four sessions was analyzed separately.It was found that the activity of amygdala on the left side of four sessions in the effective group increased in the transfer,and there was a statistical difference between the transfer activity of session 4 and that of session 3[t(6)=0.9,P=0.036],but there was no such trend in the invalid group,and there was no statistical difference between the transfer activity of session 4 and that of session 3.2.2 Results of resting-state fMRI data:ReHoDouble-sample t test was used to compare the difference of ReHo between the effective group and the ineffective group after the 3rd rtfMRI-NF training(session 3).(GRF,P<0.05).The score of ReHo maps increased after the 3rd rtfMRI-NF training were found in right midial occipital,frontal lobe,cingulate gyrus,insula,precuneus.The score of ReHo maps decreased after the 3rd rtfMRI-NF training were found in the superior frontal gyrus,right anterior cingulum,left angular,temporal lobe.Double-sample t test was used to compare the difference of ReHo between the effective group and the ineffective group after transfer training(session 4).(GRF,P<0.05).The score of ReHo maps increased after transfer training were found in the medial frontal gyrus,right anterior cingulate,inferior frontal gyrus,insula,right precentral,left superior frontal gyrus.The score of ReHo maps decreased after transfer training were found in the parietallobe,lingual gyrus,precuneus,right fusiform,parahippocampa gyrus,right angular,temporal lobe.Conclusion1 This study confirmed the feasibility of using rtfMRI-NF training to instruct ID patients to up-regulate their amygdala activity by recalling happy AMs.In particular,ID patients with good compliance and sensitivity to emotional changes are especially easy to learn the strategy of autonomic regulation of amygdala activity.2 In this study,it has been proved for the first time that rtfMRI-NF training can improve the insomnia symptoms of ID patients by enhancing the local consistency of emotion and cognitive related and improving the emotional and cognitive network of ID patients,thus improving the insomnia symptoms of ID patients.3 rtfMRI-NF training improves ID insomnia symptoms with short-term transfer effect,which provides a new idea for the clinical treatment of ID.