Table 1 Ablation protocol applied to clustering analysis.

From: A clustering approach identifies an Autism Spectrum Disorder subtype more responsive to chronic oxytocin treatment

Excluded feature

Max responder rate

Δ responder rate

Excluded feature

Max responder rate

Δ responder rate

None

61.5%

Without clustering

43.9%

17.6%

Gender

61.5%

0%

Basal saliva OXT

61.5%

0%

Age

61.5%

0%

Basal plasma OXT

61.5%

0%

ADOS-CP

61.5%

0%

Social

61.5%

0%

ADOS-SA

55.6%

−5.9%

Angry eyes

57.1%

−4.4%

ADOS-RRB

55.6%

−5.9%

Angry nose

50.0%

−11.5%

ABAS-GAC

61.5%

0%

Fear eyes

55.6%

−5.9%

SCQ

61.5%

0%

Fear nose

61.5%

0%

CSQ

61.5%

0%

Happy eyes

61.5%

0%

RBS

61.5%

0%

Happy nose

61.5%

0%

Aloof

61.5%

0%

Neutral eyes

50.0%

−11.5%

Passive

61.5%

0%

Neutral nose

50.0%

−11.5%

Active but odd

61.5%

0%

   
  1. All 23 features were used for clustering analysis. None means that none of 23 features were excluded.
  2. ADOS-CP Autism Diagnostic Observation Schedule-2-comparison score, SA social affect, RRB restrictive and repetitive behavior, ABAS-GAC adaptive behavior assessment system-II, global adaptive composite, SCQ social communication quotient, CSQ caregiver strain questionnaire, RBS repetitive behavior scale—revised, Aloof/Passive/Active but odd subscale of Beijing Autism Subtype Questionnaire, OXT oxytocin.