home • blog • mastodon • thingiverse

It's been a while since I wrote about autism. To refresh: the hypthesis is that autistic people expect the world to contain less "outliers", and themselves produce less "outliers" of behaviour. The mathematical details of this are a little subtle -- if you have a feel for the difference between a super-exponential tail and an algebraic tail you probably have the necessary mathematical background -- the rarity of such people has so far rendered the theory a memetic failure.

Anyway, a recent New Scientist article on Levy flights and search patterns got me to pondering a way to test the theory. The idea is this: Observed search behaviour results from a combination of internal strategy and external cues. If external cues are minimized, a persons's search behaviour should result purely from their internal strategy. Hopefully, autistic people's search behaviour under such conditions will have a certain statistical signature -- one that has less sudden leaps about the search space than is normal.

As an example of the kind of search task I have in mind, find the X in this figure:

 AVRSSSDDHJZMQEBMEKNCLGJICEPCLOPQZUPHRIJVNDHQYEERGQNMZWQVFRAQYDYGFVHVZYNINKCYTRZU
ZQFJHKIJMMEIGGFLLNUUGHNBOLFPJFNFFQZMEOTFPLKKDORMIVGHKRNSLFIMHSSEODYMOBJVUKVSYROO
ZLWEIZNLDQIRYTWMRCUPGBSDCYJVEUTZPVEKVGCNPCKMVFKVCKUADWSIHFCMUOQPTIFETTNBGJINNDGF
ARDROFPCNLRTCBFTSNSYOLTKTQTWVMWMTWHKWLWUGGZNWFNNSVLQFNVVHJJSGTPIJTLILHSRRCDAWIPZ
HJMHARYZPWDJIETIPDKOEVLUYUGFIZZCIQNKOQNIKEAALOZWEGPWQLNZRPKEHSIKSOEVKFMGKEQFLCIG
QJHDDRVKLPQOLLASMSEGEQNFQYBPGTUILJPVRMMGRIKAENDYMQNPTGORPQNKHDILAJVBLVTGSEYWAFAR
SZJFUDCLOVNNDDJBDLWPCTRBOHOBNVPZWNUJHUBPSCJTCVIHHJUVNKZTDPHYMYGCACRGKVCYFINUZMWE
WVRHJMJBHDAUQEJBEHCOTBKQCYSRYREJEQSGHDUTOIDCMJVQCTPAWDWDVNQKCEKAMIBZKVZAHBAGVKZP
JLDRPPFYJQFRVDARECQVWVHTDBYNVZEKNLTNVMIQUPUJZVUTPDPVMQSLKASTSPUHNVWEJSDAMYZKGGMQ
PZPUORSKOTCLYOHJZSYSNLFMHZIJHBOBLOUNYTMZZGNYIQFEWCFBULNIKPNPIAUKQGFEFJHOBUHDGFUA
OAWKBRJUMNNONGBCMHCPSCDWVVAQHNDYHQPLDLOIYCRCTBCFTSDNQPYABSIDOSTQZILFZJASIVSLPFPA
ZONCNUKHGRUHICDAVVQPUPFWQZIOVBLATAWEARLOZFTEJSVSHBESVDRWACEMLSRKDYOKRKAVOHRDZMWR
PGLUYCBDKQGMSUAYZHVELWZKMMBQMYLGDTVCFGBNEOQMQCNDFHBWKUPREIFFARNBBLPATDMCSFZDCHFE
MUKCAGAZPIFLAUFTOVOGELOPUAGGPYDDEPSPAGNRFCUQHYAGTTSOGZDWAZMUAQQVFECRUIBWJZSIFGVK
GJYSMJVPZKJEBYWRIHOVJYUHFJSFGOYIJTOYROFBSUFNZOFQPCIGQVWPCIIDLQNCONDCOJQPCIKQLTZF
SWHLSICNPCCGZLYMFPKHLVKGHYZCJAUWMYJILKBJGHPEIJKHVVSAOKWPHGHEJQAEPMTJZUJRCCEFGVMT
MLMYCNNHKDTWWYABOIHLGWLGOHSPTERRHAEDIYBSPEUDNUMOYWLBGUCNLTYZHKGHGCYFLMETDEJGTCCA
JHMWVHPMOPEMLBBZCKPCGBPRPLFRTPPVNICSWEVWMTFYEAQYLQSOTNPTYGNDIQSUWUAPKSADQBCVHJLN
TQAZWDBMIWRTTHLDWKLQKKHTBNPTDGAEPOBUOIBLFQSWNCBPWLKUKUIEOVCWWZBEUPTHGULVZKHDGDCC
OFGBZYVNAWCJRHKWRTYECSNGZDTYXYVDZEIQJWTGPILFTNGGRUSGGRZJRQLBEDFZRPOVAFIYZLGKHJQI
VWAJLHIGKWMMOUWYNDTHYPYANMNUHSFNDLLDOMHGZJAQJRKQDTWATYJZIRWPTKJDEHITZYEKTBUGYAAZ
QENSNUTJEUSNTAIMZHKAPYVESAFBCFGBZNMAWPTWVRRBFTINOBTDLAFYBMVQNWKPBPJHBVAMGFSTMABT
IUGEBNDRRCZWGFDIUUZHKQKGFJVBJVQYUQOLMDVWVUJCNAKYMAQVAPIELSAGEMHSHADLCTGUQOWZFDWD
BWNEILTHEQGLNMAIZJBFGJEVGACNIBDGSMCARDNJIAURPHQAGWJEQJKSRKYUALYVTTPNTVPHOBHOODGB
IEVIYKIRKONKHAVESLARHWLKRRZSOAGLUMDBJIDMRZVKRKNJZOKWJCVQHLYZOLRZNFVRDZVJTWVVBEIF
QYMKEBNCUEJGFLQUWJWWJECUYJAEMEHDADSIIAFNFFKMBQOBBBDPPQEGDCQWCMKLANHKNIRBEALNFIUG
EKDBKZMKGSFDDNLCCMBDCRWYIGDTMEFZTOJVBIOQTLJJEWWVTEEBKUAOEQTEUQHCOWEJUBGKOIDJVRJK
RLJGBRDSFRSELCNDCNUFGOYFTKWIQKCJLIEFLHKBIGHARSRFTYCBRALMOTOJZFCODIFBWRKELOVTSLFZ

Note how there's no trick that will let you home in on the X quickly. To the eye's pre-attentive texture discrimination, it's a more-or-less uniform field.

Observing the search path a person takes might involve eye tracking, or, in a low tech version, tracking a spotlight in a darkened room (or computer simulation thereof). I have found several papers that use eye tracking, and the signature I'm looking for is not obviously present in the eye tracks presented in these... but these papers have all involved tasks, such as watching clips from the film "Who's Afraid of Virginia Woolf", which have had a distinctly non-uniform search field.

[æ]