Skip to main content
. 2012 Jan;38(1):104–113. doi: 10.1093/schbul/sbr172

Table 2.

Task Parameters, Performance, and Reliability on 5 Versions of the DPX

Version Trial-Type Number of Trials (%) Schizophrenia Patients
Controls
% Errors (SD) α % Errors (SD) α
Long forms (ISI 4000 ms)
    Form 1 AX 88 (68.75) 9.1 (8.1) .87 3.7 (4.2) .78
AY 16 (12.5) 20.1 (17.0) .71 10.3 (10.2) .47
BX 16 (12.5) 19.9 (21.2) .83 8.3 (12.5) .74
BY 8 (6.25) 4.2 (8.5) .32 0.9 (3.7) .15
    Form 2 AX 80 (62.5) 10.2 (10.2) .90 4.3 (5.1) .80
AY 24 (18.75) 14.1 (14.6) .80 7.7 (9.2) .68
BX 16 (12.5) 17.3 (19.7) .82 9.5 (13.8) .76
BY 8 (6.25) 5.9 (9.6) .27 2.1 (6.5) .43
    Form 3 AX 76 (59.38) 10.5 (10.1) .90 5.2 (6.9) .88
AY 24 (18.75) 14.3 (14.2) .78 6.6 (8.1) .64
BX 20 (15.63) 20.5 (22.0) .85 9.0 (12.9) .80
BY 8 (6.25) 5.4 (10.3) .43 1.8 (5.0) .14
Short forms (ISI 2000 ms)
    Form 1 AX 88 (68.75) 9.1 (9.2) .90 2.8 (3.9) .80
AY 16 (12.5) 18.1 (15.1) .65 9.8 (5.0) .39
BX 16 (12.5) 20.4 (20.4) .79 7.8 (4.8) .53
BY 8 (6.25) 7.1 (11.1) .28 1.8 (7.4) .21
    Form 2 AX 80 (62.5) 9.6 (9.9) .90 3.2 (12.3) .84
AY 24 (18.75) 14.2 (13.1) .74 6.2 (6.4) .58
BX 16 (12.5) 20.2 (19.8) .80 8.2 (5.0) .73
BY 8 (6.25) 7.1 (10.3) .27 2.1 (4.8) .38

Note: ISI, interstimulus interval (time between the offset of the cue and the onset of the probe, in milliseconds); Reliability is measured by Cronbach’s α for available participants; For all versions of the task, the number of trials was 128, cue length was 1000 ms, probe length was 500 ms, and intertrial interval was 1200 ms. Long form 1 represented the standard version of the task. The other task versions were developed by altering the standard version’s ISI and critical trial (AY or BX) frequencies. Long forms 2 and 3 were created by increasing the frequency of one (long form 2; AY trials) or both (long form 3; AY and BX trials) critical trials. Short form 1 was created by reducing the ISI, whereas short form 2 reduced ISI and increased AY trial frequency. AY or BX trial frequency increases were accommodated by a reduction in AX trial frequency (ie, the expectancy manipulation).