Insomnia: a promising cure

I once saw a talk by An­ders Sand­berg in which he said the best cog­ni­tive en­hancers are caf­feine, modafinil and sleep. The former two have diminish­ing re­turns, but the lat­ter does not. It’s pretty clear that sleep is a ma­jor de­ter­mi­nant of pro­duc­tivity and mood. I have no­ticed that a num­ber of EAs suffer from in­som­nia and are search­ing for a vi­able cure, so I thought I would out­line one that has worked in­cred­ibly well for me, and is well-sup­ported by ev­i­dence. This is in the hope of sub­stan­tially im­prov­ing the pro­duc­tivity and mood of EAs, which I think is a very high value thing. This is a post about the best nootropic go­ing (for in­som­ni­acs): CBT for in­som­nia, in­clud­ing sleep re­stric­tion.

Per­sonal experience

I suffered from mod­er­ate to se­vere in­som­nia through­out my twen­ties, and had par­tic­u­larly se­vere in­som­nia for a cou­ple of years be­fore I turned 30. I was pre­scribed sleep­ing pills which don’t work or have bad side-effects. I have learned that this is pretty typ­i­cal treat­ment in the UK. Dur­ing the nadir, I would strug­gle to get to sleep for two hours, then sleep for four hours, then wake up at 5am and be un­able to get back to sleep, ly­ing in bed un­til 9am. I tried var­i­ous sleep­ing pills for this, in­clud­ing mela­tonin (way too much, it turns out), which helped a bit.

I then tried CBT-in­som­nia in­clud­ing sleep re­stric­tion us­ing an on­line course (called Shut-i which now ap­pears not to be available for some rea­son) and was cured af­ter about 2 months and have been ever since. I now sleep up­wards of 8 hours ev­ery night. This ap­proach has been similarly suc­cess­ful for a friend suffer­ing se­vere in­som­nia who tried var­i­ous pre­scribed sleep­ing pills with lit­tle suc­cess.

Ev­i­dence for CBT for insomnia

Anec­dotes aside, CBT for in­som­nia is well-sup­ported by ev­i­dence. For ex­am­ple, see these two sys­tem­atic re­views and this re­view by the Amer­i­can Academy of Sleep Medicine. NICE, the re­spected util­i­tar­i­an­ish body that de­ter­mines health­care pri­ori­ti­sa­tion in the UK says that:

  • There is good ev­i­dence that stim­u­lus con­trol ther­apy, re­lax­ation ther­apy, and cog­ni­tive be­havi­oural ther­apy (CBT) are effec­tive in the treat­ment of long-term in­som­nia [Mor­gen­thaler et al, 2006; Wil­son et al, 2010].

  • There is mod­er­ate ev­i­dence that sleep re­stric­tion, para­dox­i­cal in­ten­tion, and biofeed­back are effec­tive in the treat­ment of long-term in­som­nia [Mor­gen­thaler et al, 2006].

Up­date: Luke Muehlhauser has ques­tioned the strength of the ev­i­dence on CBT-I in a com­ment be­low. Note that he does still recom­mend CBT-I to peo­ple suffer­ing in­som­nia, given the limited costs as­so­ci­ated with CBT-I. I dis­agree with Luke’s read­ing of the ev­i­dence, as I dis­cuss in the com­ment.

Pre­vailing treat­ment for insomnia

In spite of the ev­i­dence, in­som­nia is very poorly treated, in the UK at least, with doc­tors of­ten recom­mend­ing sleep­ing pills that don’t work and have bad side-effects. I know peo­ple who have tried these with­out suc­cess and were told that they had run out of op­tions. Given that CBT for in­som­nia is recom­mended by NICE, this is a pretty crazy situ­a­tion. More­over, the treat­ment is non-phar­ma­colog­i­cal and doesn’t re­quire a trained ther­a­pist.

What does CBT for in­som­nia in­volve?

I recom­mend Over­com­ing In­som­nia by Colin Espie for a good and short out­line.

  • Sleep re­stric­tion. Per­son­ally, I found this the most effec­tive as­pect. Peo­ple with in­som­nia tend to spend way too long in bed in the be­lief that this will in­crease their chance of sleep. This is the op­po­site of what you should do.

    • Stay in bed for at most as long as the to­tal amount of sleep you ideally need. If you need 8 hours of sleep, spend at most 8hrs 30mins in bed, and no more.

    • Don’t read or watch TV in bed, just sleep/​have sex.

    • In­som­nia suffer­ers should aim to more strictly re­strict their time in bed. If you are sleep­ing five hours a night, you should aim to grad­u­ally limit the time you spent in bed, down from 7 hours to 5 hours. This re­duces the psy­cholog­i­cal as­so­ci­a­tion be­tween bed and lack of sleep, and so re­duces the stress re­sponse you ex­pe­rience when in bed. Once you have bro­ken the as­so­ci­a­tion, you can scale your time in bed back up. A full plan is out­lined in the Espie book.

  • Sleep hy­giene: out­lined at this NHS link, and this NSF link.

  • Not­ing down and ra­tio­nally deal­ing with wor­ries, a la CBT, also men­tioned in the NHS link.

  • Dim­ming the lights: A form of sleep hy­giene that de­serves a spe­cial men­tion. Many EAs wan­der round in or­ange gog­gles af­ter 9pm, un­cowed by so­cial mores. This is a good thing—I sug­gest dim­ming the lights through­out your house, in­clud­ing the bath­room, about 2 hours be­fore you go to bed.

NB. I am eter­nally grate­ful to Daria Belos­tot­skaya for sug­gest­ing this solu­tion to me.