2011. Vol.2, No.7, 743-753
Copyright © 2011 SciRes. DOI:10.4236/psych.2011.27114
Do Auditory Temporal Discrimination Tasks Measure Temporal
Resolution of the CNS?
Ian T. Zajac, Nicholas R. Burns
School of Psychology, University of Adelaide, Adelaide, South Australia.
Received July 1st, 2011; revised August 5th, 2011; accepted September 15th, 2011.
Rammsayer & Brandler (2002) have proposed that auditory temporal discrimination tasks provide a measure of
temporal resolution of the CNS which is argued to be partly responsible for higher order cognitive functioning.
We report on two studies designed to elicit the nature of the functions underpinning these auditory tasks. Study 1
assessed whether temporal generalisation (TG) might be better considered as a measure of working memory
rather than of temporal resolution of the CNS. In N = 66 undergraduates TG did not predict speed of processing
tasks; however, there was evidence of a relationship between TG and working memory. Study 2 reanalyzed pre-
viously published data on temporal discrimination tasks and showed that the relationship between auditory tem-
poral tasks and intelligence reflects memory functions and processing speed. Auditory temporal discrimination
tasks are confounded by speed and memory and should not be considered as measures of temporal resolution of
the CNS.
Keywords: Temporal Discrimination, Working Memory, Intelligence, Auditory Reaction Time
The last few decades have seen a shift in focus from the
taxonomic study of cognitive abilities to the identification of
lower-order cognitive and physiological correlates of human
intelligence (Neubauer & Fink, 2005). This shift has been
driven by a desire to identify the biological roots of higher or-
der cognition (Stankov, 2005). The exploration of biological
correlates of intelligence has been aided by advances in the
measurement of brain activity. Studies employing electroen-
cephalograms, for instance, have reported that peripheral nerve
conduction velocity and event related potentials share variance
with cognitive ability measures (Burns, Nettelbeck, & Cooper,
2000; Reed & Jensen, 1993). On the other hand, in order to
measure lower-order cognitive processes, researchers have
turned to a class of tasks termed Elementary Cognitive Tasks
(ECTs). The impetus for this is that ECTs are characteristically
easy tasks which putatively rely on a limited number of mental
processes or operations (Carroll, 1993). Thus, they supposedly
provide cleaner measures of biological processes than tradi-
tional, more complex tests (Stankov, 2005).
The two most commonly researched ECTs are reaction time
(RT) and inspection time (IT): RT tasks measure the speed with
which an individual is able to respond to a particular reaction
stimulus; and IT tasks measure the minimum exposure duration
required to accurately discriminate stimuli that differ on some
dimension. Both classes of tasks are held to reflect information
processing speed (Jensen, 2005). However, it has been found
that performance in these tasks is relatively independent. The
correlation between these ECTs is seldom more than r = .30,
with the strength of the correlation appearing to increase as
complexity of the RT task increases (Burns & Nettelbeck, 2003;
O’Connor & Burns, 2003; Petrill, Luo, Thompson, & Detter-
man, 2001).
Despite their relative independence, IT and RT tasks have
been found to share a statistically significant amount of vari-
ance with measures of psychometric intelligence. People with
higher speed of information processing—faster average RTs
and shorter ITs—perform better on tests of cognitive ability
than those who are slower. It has been proposed that RT and IT
could account for as much as 25% of the variance in intelli-
gence test performance (Grudnik & Kranzler, 2001; Jensen,
1982, 2005, 2006; Nettelbeck, 1987, 2001, 2003). However, a
more recent meta analysis which based its conclusions on 1146
correlations between speed of processing measures and intelli-
gence measures proposes a much smaller effect: around 10%
shared variance for RT and intelligence and about 8% between
IT and intelligence (Sheppard & Vernon, 2008).
Regardless of the size of these effects, and returning to the
idea of identifying the biological basis of intellectual function-
ing, it is necessary to explain the observed relationship between
intelligence and performance in ECTs. Many explanatory mod-
els appeal to the concept of “neural efficiency” as the determi-
nant of both information processing speed and intelligence (see
e.g., Hendrickson, 1982; Hendrickson, 1982; Vernon, 1993).
Jensen’s (1982) model of neural oscillations, for example, pro-
ceeds from the assumption that RT provides an index of the
efficiency of the central nervous system (CNS). Individual
differences in both processing speed—as measured by ECTs—
and intellectual functioning are attributed to differences in the
rate of oscillation between refractory and excitatory states of
neurons. The transmission of neurally encoded information is
assumed to be more efficient as well as faster at a higher rate of
neural oscillations. This is because it takes less time for a neu-
ron to re-enter its excitatory phase when processing information
than when oscillations are slow.
An alternative theory linking higher-order cognitive proc-
esses to elementary functions has recently been revisited by
Rammsayer and others (Helmbold & Rammsayer, 2006;
Helmbold, Troche, & Rammsayer, 2006a, 2007a; Rammsayer
& Brandler, 2002, 2004, 2007). Originally proposed by Sur-
willo (1968), this theory also appeals to a hypothetical oscilla-
tory, or “clock”, mechanism in the CNS to explain individual
differences in speed of information processing and intelligence.
Thus, “if the hypothesised internal master clock of individual A
works at half the clock rate as the one of individual B, then A
does not only need twice as long as B to perform a specific
sequence of mental operations, but also the occurrence prob-
ability of interfering incidents will be increased” (Rammsayer
& Brandler, 2007: p. 124); according to the theory this results
in both slower performance on speed of processing tasks and
lower intelligence. The central features of the internal clock
mechanism are a pacemaker and an accumulator (Rammsayer
& Brandler, 2002).
Jensen (2006) cites findings from ECT research to support
his neural oscillation model. In order to obtain empirical sup-
port for the master clock theory, Rammsayer and colleagues
have sought to demonstrate that presumed measures of clock
rate differ between individuals of low and high intelligence
(Helmbold et al., 2006a, 2007a; Rammsayer & Brandler, 2002,
2004, 2007). They have argued that accuracy on psychophysi-
cal timing tasks—by analogy with performances on ECTs—
reflects basic processes related to neural efficiency (Helmbold,
Troche, & Rammsayer, 2007b). According to this theory the
number of neural oscillations generated by the pacemaker dur-
ing a timed interval is recorded by the accumulator and be-
comes the internal representation of that interval. Thus, the
higher the frequency of oscillations the finer the temporal reso-
Because audition has finer temporal resolution than vision
(Rammsayer & Brandler, 2002), attempts to measure temporal
resolution of the CNS have focussed on auditory tasks. Ramm-
sayer and Brandler (2002) found that auditory duration dis-
crimination was significantly better for a high-IQ group than
for a low-IQ group and that it explained around 20% of the total
variance of a single fluid intelligence (Gf) measure. A later
study concluded that a general pacemaker based interval timing
mechanism is involved in auditory temporal order judgement,
duration discrimination, and temporal generalisation and that
performances on these tasks is independent of general auditory
discrimination ability (Rammsayer & Brandler, 2004). Factor
scores on this general timing (Gt) mechanism have subse-
quently been shown to share substantial variance (about 25%)
with psychometric measures of general intellectual ability, oth-
erwise referred to as “g” (Rammsayer & Brandler, 2007).
Whether Gt solely reflects temporal resolution is arguable
and Helmbold et al. (2006a) have explored whether sensory
discrimination abilities rather than temporal resolution of the
CNS account for the relationship between Gt and intelligence.
Temporal generalisation and pitch discrimination performance
was measured and regression analyses showed these tasks
combined to predict 25% of the variance in g factor scores. The
unique contributions of temporal and pitch tasks were 9% and
6%, respectively. The shared, and presumably general, sensory
processes accounted for the remaining 10% of predicted g.
Helmbold et al. (2006a) concluded that the unique contribution
of temporal discrimination to the prediction of g supports the
notion that it measures specific aspects of neuronal information
processing related to intellectual capacity but independent of
non-temporal aspects of sensory discrimination.
As noted, it has been proposed that auditory psychophysical
timing tasks are analogous to existing ECTs in terms of meas-
uring basic processes related to neural efficiency (Helmbold et
al., 2007b). Therefore, one should expect these tasks to corre-
late at least moderately with existing ECTs, including RT and
IT, but evidence regarding this hypothesis is equivocal (Helm-
bold et al., 2007a; Rammsayer & Brandler, 2007). Although the
correlation between latent RT and latent temporal discrimina-
tion factors appears moderately strong (r = .65: Helmbold et al.,
2007b), the correlation between individual temporal tasks and
RT parameters is markedly weaker and in many cases not sta-
tistically significant. The average correlation between eight
temporal tasks and different RT parameters in Helmbold and
Rammsayer (2006) was only r = –.19 (SDr = .07)1. This ab-
sence of significant correlations between temporal tasks and RT
suggests that temporal discrimination tasks may not be meas-
uring the elementary processes reflected in RT tasks.
Nonetheless, these findings have been interpreted as provid-
ing evidence that auditory temporal discrimination tasks index
temporal resolution in the CNS. There are, however, several
issues with these tasks which question whether the observed
correlation between temporal performance and intelligence is a
result of neural efficiency and, by extension, temporal resolu-
tion of the CNS.
First, although it has been supposed that these tasks are ele-
mentary (Rammsayer & Brandler, 2004) some do not appear to
be. The Temporal Generalisation (TG) task appears rather more
complex than archetypal ECTs. TG requires participants to
judge whether a test stimulus is the same as a standard stimulus
learnt in a pre-exposure phase. Thus, the task requires: 1) accu-
rate learning of the standard stimulus; 2) accurate registration
of the test stimulus; 3) accurate retrieval of the learnt standard;
and 4) a successful comparison of the test and learnt standard,
in order to complete each test item. It appears that the cognitive
operations required in this test are complex, and even if tempo-
ral resolution of the brain itself is independent of higher order
cognitive operations (Rammsayer & Brandler, 2002: p. 509),
performance on this task is not likely to be. It is plausible that
the observed relationship between TG and intelligence reflects
the shared cognitive operations common to TG and general
intelligence tests, rather than temporal resolution of the brain.
This hypothesis applies to other discrimination tasks used in
these studies. Duration Discrimination (DD), for example, re-
quires participants to compare two successively presented time
intervals to decide which was longer. Thus, an internal repre-
sentation of each interval must be formed and, given the length
of the intervals—1 sec or longer in one condition—and the ISI
(900 ms), these representations need to be accessible for up to
three seconds after presentation; at least for the first presented
interval. Unless the accumulator in the master clock theory
incorporates an information storage component, performance
on this task is also likely to rely on complex cognitive functions.
Considering the requirements of both TG and DD, it appears
that the cognitive operations involved may include substantial
memory functions.
A second issue regarding these findings is that most have
focused on factor scores. As such, little information is gleaned
in terms of the relationship between specific temporal dis-
crimination tests and intelligence measures. Moreover, the na-
ture of the latent construct defined by the temporal tasks is
merely surmised based on theory of what the tests have in
common. In order to accurately assess whether internal clock
rate—or temporal resolution of the CNS—is related to intellec-
tual capacity, it must first be established that each of the indi-
vidual tests used provides some measure of this. Only then can
a latent variable defined by these tests be taken to represent the
internal clock rate.
This paper presents the findings of two studies on temporal
discrimination tasks. The impetus for these studies was to in-
vestigate whether temporal discrimination tasks provide a
measure of elementary functions such as temporal resolution of
1These correlations were not reported in the published article and we are
grateful to Prof. Rammsayer for providing these to us.
I. T. ZAJAC ET AL. 745
the CNS, or whether they might better be conceptualized as
measures of more complex cognitive operations like memory
Study 1
The Temporal Generalisation (TG) task has been shown to
relate to g but its relationship with specific cognitive abilities
and ECTs has not been considered. Thus, the relationship may
reflect executive cognitive functions utilised in task perform-
ance and not neural efficiency, as proposed. The purpose of this
study was to provide a test of this hypothesis by exploring the
relationship between TG and measures of processing speed (Gs)
and working memory (WM). Importantly, speed of processing
was measured by traditional speed tasks and ECTs, including
RT and IT, because of the considerable evidence that these
ECTs are reliable measures of elementary functions (Jensen,
2006; Nettelbeck, 2001). If TG measures elementary functions
as opposed to executive cognitive functions, then the relation-
ship between TG and Gs will be stronger than that between TG
and WM.
We used the same dissociation paradigm as Helmbold et al.
(2006a). The purpose was to assess the direct relationship of
TG to Gs after partialling out variance due to general sensory
discrimination processes; reflected in the pitch discrimination
task (APd). If TG measures elementary processes related to
intelligence and which are independent of general sensory dis-
crimination, then TG should make a direct contribution to the
prediction of intelligence test performance.
The superior temporal discrimination in audition has been the
motivation for the use of auditory tasks. However, if the master
clock which determines performance on these tasks is a general
feature of the neural system, it should also be responsible for
temporal discrimination in other modalities. Therefore, the
current study sought also to measure temporal resolution of
vision. This was achieved through adapting the dissociation
paradigm for the visual system to include a temporal and a
line-length discrimination task. Line-length discrimination abil-
ity would be measured to assess variance in visual temporal
performance reflecting general sensory processing. The correla-
tion between the visual and auditory temporal discrimination
tasks should be at least moderately strong if they reflect the
same elementary timing processes.
Rammsayer and Brandler (2002) have reported that temporal
resolution of the CNS is independent of cognitive operations.
We tested this assumption by introducing a backward masking
condition for the discrimination tasks. Masking has previously
been used in visual and auditory modalities to investigate tem-
poral processes underpinning perception, and which operate at
a precognitive level (Breitmeyer, 2007). If temporal tasks
measure temporal resolution of the CNS then their relationship
with cognitive ability measures should not be negatively af-
fected by the introduction of a masking stimulus; because it
emphasises pre-cognitive functioning. In fact, the strength of
the relationships might be expected to increase.
The Participants were N = 66 undergraduate students of the
University of Adelaide, South Australia. There were 7 males
and 26 females in each of the masked and unmasked conditions.
All participated as part of their Level I Psychology course re-
The presentation of all tasks and recording of responses was
controlled by one of two identical computers. Visual stimuli
were presented on 17 inch LCDs. Auditory stimuli were pre-
sented via Sony MDR-XD100 stereo headphones. Auditory
tones were calibrated prior to the study using a Radio Shack 33 -
4050 Sound Level Meter.
Discrimination Tasks
Auditory and visual discrimination abilities were assessed
using the experimental dissociation paradigm developed by
Gibbons, Brandler, & Rammsayer (2002); stimuli varied on
two dimensions simultaneously. The first dimension was tem-
poral: there were seven levels of stimulus duration. The second
dimension for the auditory modality was pitch and for the vis-
ual modality was line-length; there were seven levels of each
(see Appendix A). Line length dimensions were piloted on a
small number of colleagues to be at a comparable level of dif-
ficulty to the duration levels.
The design of the set of stimuli for the dissociation paradigm
is based on the requirement that: 1) for duration, as well as
pitch/line length, there should be a probability for the standard
stimulus of .33 in the total number of trials; 2) within each level
of one stimulus dimension, each level of the other dimension
should be represented; and 3) for each of the seven levels of
one stimulus dimension, there should be a probability of .33 for
the occurrence of the standard of the other stimulus dimension.
Simultaneous variation on two dimensions according to these
requirements results in a set of 81 stimuli for each of the visual
and auditory tasks, resulting in the frequency distribution pre-
sented in Appendix A. The test phase for each of the discrimi-
nation tasks comprised 81 trials, including 27 presentations of
the standard and nine presentations of each nonstandard stimu-
lus. Presentation order within each task was pseudo-randomised,
with the restriction that there were no more than two successive
presentations of the standard. The outcome measure for each of
the discrimination tasks was percentage of standard stimuli
correctly identified.
Auditory Temporal and Pitch Discrimination Tasks. In each
task, participants were required to identify the standard tone
among the set of nonstandard tones. Participants were in-
structed to attend solely to tone duration in the temporal task,
and to tone frequency in the pitch task. All tones were pre-
sented at an intensity of 67 db. Each task was preceded by a
learning phase in which participants were asked to learn the
standard tone. For the temporal task, a standard tone duration
(i.e., 200 ms) with a pitch (900 Hz) not administered during the
test phase was presented five times. For the pitch task, the
learning phase consisted of five presentations of the standard
tone (i.e., 1000 Hz) for 260 ms, a duration which was not in-
cluded in the test period. The testing phase immediately fol-
lowed and the onset of each trial was marked by the presenta-
tion of a visual fixation point (small white cross) in the centre
of the computer screen. After a foreperiod of 1000 ms the trial
stimulus was presented and the cross remained on the screen. In
the masking condition, a burst of white noise immediately fol-
lowed the trial stimulus for 500 ms, otherwise the trial termi-
nated. Following each trial the participant mouse-clicked one of
the onscreen buttons (“standard” or “nonstandard”) to indicate
whether they thought the trial stimulus matched the frequency
or duration of the standard tone, depending upon which task
was being completed. Feedback was given for each trial in the
form of a “correct” or “incorrect” on-screen message which
was displayed for 500 ms. Subsequent trials commenced im-
mediately after the feedback.
Visual Temporal and Line-Length Discrimination Tasks. The
requirements of these tasks were similar to the auditory tasks.
White horizontal lines presented against a black computer
screen were used analogously to tones in the auditory tasks. For
the temporal task, participants were asked to attend solely to
stimulus duration whilst in the line-length task they were asked
to attend solely to the length of the line. The learning phase for
the temporal task consisted of five presentations of the standard
duration (i.e., 200 ms) with a line length (6 cm) not adminis-
tered during the test period and for the line-length task, con-
sisted of five presentations of a standard 10 cm line for a dura-
tion (260 ms) not included in the test. The testing phase imme-
diately followed and each trial was marked by the onset of a
visual fixation point (small white cross) in the centre of the
computer screen. After a foreperiod of 1000 ms the visual fixa-
tion point was replaced by the trial stimulus. In the masking
condition, a 4 × 8 grid of 16 cm wide by 6 cm high lines imme-
diately followed the trial stimulus for 500 ms (see Figure 1)
otherwise the trial terminated. The response format was the
same as for the auditory tasks with participants indicating
whether they thought the test stimulus matched the duration or
line-length of the standard.
Working Memory Task
Dot Matrix Test (DM). A computer-administered version of
the Dot Matrix Test (Law et al., 1995) was used as a measure of
working memory (WM). Participants verified a series of simple
matrix equations whilst simultaneously remembering the loca-
tions of dots on a 5 × 5 grid. Matrix equations were either addi-
tion or subtraction equations presented as lines drawn on 3 × 3
dot matrices. Participants verified each equation by mouse-
clicking either the “True” or “False” buttons displayed on the
screen within 10 seconds, otherwise they received a prompt
(“response required”). Following an incorrect response the
message “No, look again closely” was displayed, and the equa-
tion remained until a correct was response was given.
Following correct responses a 5 × 5 grid was displayed for
1500 ms with a dot presented in one of the squares. There were
four levels during the test (2, 3, 4, and 5 equation-grid pairs
each with 4 items) and this equation-grid sequence was re-
peated according to the level. At the end of each equation-grid
sequence, a blank 5 × 5 grid was displayed on the screen. Par-
ticipants were required to mouse-click the spaces on the blank
grid which had contained the dots during the trial sequence.
Participants could not select more grid spaces than there were
equation-grid pairs but they could select fewer grid spaces (e.g.,
3 of 5 dot locations). An “enter” button was clicked after loca-
tions were selected. Three practice questions consisting of two
equation-grid pairs preceded the test. The measure for the task
was the number of dot positions, out of a total of 56, correctly
Speed of Processing
Symbol Digit (SD). A computerised coding task was em-
Figure 1.
Target and masking stimuli u sed in the visual discrimination tasks.
ployed as a measure of Gs (see McPherson & Burns, 2005, for
a detailed description of this task). A code table was presented
at the top of the computer screen throughout the task. This
comprised of nine symbols arranged horizontally, to which nine
digits were paired; digits were presented directly beneath the
symbols so that they were aligned. For each item, one symbol
was presented in the centre of the computer screen and partici-
pants responded by left clicking the mouse on its corresponding
digit in a 3 × 3 numerical grid positioned at the bottom of the
screen. Subsequent items did not commence until a correct
response was registered. Participants were required to complete
two practice trials correctly before they proceeded to the test.
The outcome measure was the number of items correctly com-
pleted in 120 seconds.
Audio Code (AC). This task was developed in our laboratory
to be an auditory analogue of the symbol digit task described
above. It has good reliability (r = .89) and correlates well with
other speed measures. In this task, a code table is displayed at
the top of the computer screen for the duration of the task. This
comprised of pictures of eight musical instruments arranged
horizontally, to which one of the numbers one through eight
was paired; the instruments were a snare, trumpet, guitar, cym-
bals, piano, bell, harp and violin. For each item, the sound of
one of the instruments was presented via headphones at an in-
tensity of 65 db. Participants responded by left clicking the
mouse on its corresponding digit in a 2 × 4 numerical response
grid positioned at the bottom of the screen. Subsequent items
commenced after a response was registered. Participants were
required to complete four practice trials correctly before they
proceeded to the test. The outcome measure was the number of
items correctly completed in 120 seconds.
Visual Inspection Time (VIT). Stimuli were presented on a
video monitor at a viewing distance of approximately 60 cm.
Preceding the target figure was a warning cue of approximately
520 ms; the cue was a small white plus (+) sign measuring 6 ×
6 mm, presented in the centre of the computer screen. The tar-
get figure consisted of two vertical lines; one measured 15 mm
and the other 30 mm. These were joined at the top by a hori-
zontal line of approximately 18 mm. A “flash mask” (see Evans
& Nettelbeck, 1993) of 375 ms immediately replaced the target
figure and consisted of two vertical lines 35 mm in length,
shaped as lightning bolts. The shorter line appeared on either
side of the target figure equiprobably.
A computerised tutorial preceded the test phase and the in-
structions emphasised accuracy rather than speed of responding.
What was required was explained using diagrams, along with
unmasked target stimuli. Practice trials required 10 correct
trials out of 10 with a stimulus onset asynchrony (SOA) of
approximately 835 ms; 10 correct trials out of 10 with SOA
approximately 420 ms; and nine correct trials out of 10 with
SOA approximately 250 ms. The estimation process began with
SOA approximately 250 ms and followed an adaptive staircase
algorithm (Wetherill & Levitt, 1965). The algorithm required
three correct responses at any SOA before SOA was reduced by
approximately 17 ms. The average SOA was calculated over
eight reversals of direction on the staircase, giving an estimate
of the SOA with an associated probability of 79% of making a
correct response. Participants indicated on which side the short
line appeared by clicking either the left or right mouse button,
Auditory Reaction Time (ART). This task required partici-
pants to respond as quickly as possible to an auditory target
stimulus. To start each trial, the participant pressed the number
“5” key in the numeric keypad on the computer keyboard. After
I. T. ZAJAC ET AL. 747
300 ms a short beep (100 ms at 880 Hz) was presented to con-
firm the trial had started. The target stimulus was then pre-
sented after a silent interval of variable duration (1300 ms,
1700 ms, 2100 ms, 2500 ms), and it was a 500 ms “bell” sound
centered on a frequency of 800 Hz. Participants were instructed
to lift their finger off the number “5” key when they heard the
target sound and press the number “8” key in the numeric key-
pad as fast and as accurately as possible. The test phase con-
sisted of 32 trials before which participants had to complete
five practice trials correctly. Mean RT was calculated after
removing errors and outliers (±3 SD). The average number of
trials remaining after these removals—and from which Mean
RT was derived—was M = 31.30 (SD = .63, Min = 30, Max =
Upon attending the testing session participants were assigned
to either the masked or unmasked condition depending on
whether they were an odd or even numbered participant. They
were seated in a quiet room in the laboratory and were guided
through the tasks by the computer. The four discrimination
tasks (see below) were interspersed with cognitive ability
measures, which were ordered as they are set out below. The
discrimination tasks were ordered so as to switch between mo-
dalities (auditory pitch/visual length/auditory temporal/visual
temporal), and the discrimination tasks were counterbalanced
within conditions to reduce fatigue effects (visual length/audi-
tory pitch/visual temporal/auditory temporal). The ordering of
cognitive ability measures remained constant. The testing ses-
sion took 60 minutes to complete.
After collating the data it was apparent that two participants
did not complete Dot Matrix (DM), one participant failed to
complete Auditory Temporal discrimination (ATd) and another
participant failed to complete Symbol Digit (SD). These miss-
ing data were replaced using the Expectation Maximization
(EM) method in Missing Values Analysis in SPSS v.15. Fol-
lowing this an outlier analysis was performed by standardizing
scores on each variable. The only identified outlier was for
Audio Code (AC; z = 3.14), which was deleted and subse-
quently replaced using EM.
Descriptive statistics for the cognitive measures and dis-
crimination tasks are presented in Table 1. As can be seen,
performance in the masked condition was poorer for all of the
discrimination tasks, with small to large effects. The difference
was only statistically significant for the auditory temporal dis-
crimination task (ATd [t(64) = 2.11, p = .038]) and auditory
pitch discrimination task (APd [t(64) = 2.77, p = .007]).
Table 2 presents the correlations between the cognitive tests
for the total sample and Table 3 displays the correlations be-
tween the discrimination tasks for the masked and un-masked
conditions. As can be seen the correlations between the cogni-
tive tests are small-to-moderate and the correlations between
the discrimination tasks are moderate-to-strong. Of particular
note is the correlation between ATd and Visual Temporal (VTd)
discrimination. As hypothesized, the correlation between them
is notably strong indicating that to a large degree these tasks
index the same construct.
In order to assess the extent to which the temporal tasks pre-
dict performance in the speed tasks and working memory task
(DM), linear regression was used. Rather than regress each of
the speed tasks onto the discrimination tasks, independently, a
composite speed measure was calculated by averaging stan-
Table 1.
Descriptive statistics for discrimination tasks, cognitive measures, VIT
and RT.
a SD Min Max db
Unmasked.58 .18 .07 .89
Masked .52 .19 .15 .89
Unmasked.68 .16 .26 .93
Masked .58 .19 .11 .89
Unmasked.81 .13 .52 1
Masked .77 .11 .56 .96
Unmasked.64 .15 .26 .89
Masked .53 .16 .19 .81
SD 90.8 16.1 64 133
AC 63.8 7.9 49 82
DM 38 6.3 19 51
VIT (ms) 45.3 11.6 19.5 76.4
ART (ms) 502.6 115.4 312.4 768.1
VTd = Visual Temporal Discrimination; VLd = Visual Length Discrimination;
APd = Auditory Pitch Discrimination; ATd = Auditory Temporal Discrimination;
SD = Symbol Digit; AC = Audio Code; DM = Dot matrix; VIT = Visual Inspec-
tion Time; ART = Auditory Reaction Time. aTemporal tasks = percent correct;
SD, AC & DM = N correct;VIT and RT = msec, bCohen’s d.
Table 2.
Correlations between discrimination tasks for masked (above diagonal)
and unmasked conditions (below diagonal).
VTd - .36* .65** .27
VLd .34* - .35* .17
ATd .64** .20 - .50**
APd .59** .30* .66** -
VTd = Visual Temporal Discrimination; VLd = Visual Length Discrimination;
APd = Auditory Pitch Discrimination; ATd = Auditory Temporal Discrimination
*p < .05 (1-tailed) **p < .01 (1-tailed)
Table 3.
Correlations between cognitive tests.
AC .47**
DM .25* .34**
VIT –.17 –.15 –.24*
ART –.39** –.21* –.20 .00
SD = Symbol Digit; AC = Audio Code; DM = Dot matrix; VIT = Visual Inspec-
tion Time; ART = Auditory Reaction Time. *p < .05 (1-tailed) **p < .01 (1-tailed).
dardized scores on these variables (SD, AC, VIT, ART). A
series of models were subsequently run in which either the
composite speed measure or working memory measure (DM)
was the dependent variable. The visual discrimination or audi-
tory discrimination tasks were used as independent/predictor
The results of these analyses are presented in Table 4. As can
be seen, none of the models was statistically significant. The
association between discrimination tasks and DM does however
appear to be stronger than for the composite speed measure as
well as more consistent. It is of a comparable magnitude in each
of the modalities and in the different masking conditions. Be-
cause the discrimination tasks are effectively identical in both
conditions—they differed only in terms of the addition of a
backward-masking stimulus—the regressions with DM as the
dependent variable were repeated using the total sample. Ac-
cording to these models, visual discrimination tasks and audi-
tory discrimination tasks predicted a statistically significant
amount of variance in DM (visual model [R² = .09, F(2, 63) =
3.26, p = .045] and auditory model [R² = .11, F(2, 63) = 3.74, p
= .029]), and the sizes of the effects remained consistent with
those in Table 5. The standardized coefficients for the auditory
temporal and pitch tasks were β = .30 and β = .04, and β = .22
and β = .15 for the visual temporal and line length tasks. Thus,
in both modalities the temporal task is the stronger predictor of
Discussion—Study 1
The relationship between Temporal Generalisation (TG) and
markers of specific cognitive abilities was explored. The
analyses suggest TG relates more strongly to the marker of
Working Memory (WM) than to the composite speed measure.
This result provides only limited support for the hypothesis that
TG measures executive cognitive functions and not temporal
resolution of the CNS because of the lack of statistical power
and the limited number of marker tests.
Study 2
In light of the limited evidence provided in Study 1, the pur-
pose of the current study was to explore further whether tem-
poral tasks rely on memory functions by reanalysing previously
published data. Rammsayer and Brandler (2007) reported on
five temporal discrimination tasks; the Hick RT task (Hick,
1952); and a well defined battery of cognitive ability tasks
measuring different aspects of intelligence corresponding to
Thurstone’s (1938) primary mental abilities. These tasks were
Table 4.
Regression models for masked and unmasked conditions.
DV IV Condition R² F df p
Masked .02 .22 2, 300.80
VTd & VLd
Unmaksed.02 .23 2, 300.80
Masked .02 .33 2, 300.72
ATd & APd
Unmasked.09 1.51 2, 300.24
Masked .12 1.96 2, 300.16
VTd & VLd
Unmaksed.08 1.33 2, 300.28
Masked .09 1.45 2, 300.25
Dot Matrix
ATd & APd
Unmasked.10 1.73 2, 300.20
VTd = Visual Temporal Discrimination; VLd = Visual Length Discrimination;
APd = Auditory Pitch Discrimination; ATd = Auditory Temporal Discrimination
DV = Dependent Variable; IV = Independent Variable.
completed by a large sample (N = 100). The temporal tasks
included: 1) Duration Discrimination (DD), requiring a decision
concerning which of two successively presented timed intervals
was longer; 2) Rhythm Perception (RP), requiring a decision
concerning which of five beat-to-beat silent intervals—marked
by 3 ms clicks—deviated from the constant 150 ms duration; 3)
Temporal-order Judgment (TOJ), in which participants decide
whether the onset of a Visual LED preceded that of an auditory
stimulus, or vice versa; 4) Auditory Flutter Fusion (AFF),
which derives an estimates of the ISI at which two successively
presented auditory noise bursts appear fused; and 5) Temporal
Generalisation (TG). These tasks comprise the battery used in
previous investigations of temporal discrimination (see Helm-
bold & Rammsayer, 2006; Helmbold et al., 2007b; Rammsayer
& Brandler, 2002, 2004, 2007).
Rammsayer and Brandler (2007) reported that a temporal g
(Gt) factor defined by the discrimination tasks predicted 31% of
variance in psychometric g, as defined by the cognitive ability
measures. Combining Gt and a Hick g factor increased the
proportion of explained psychometric g by only 2%. The
unique contribution of temporal g was 20.5%, the shared con-
tribution of temporal and Hick g was 10.5%, and the unique
contribution of Hick g was only 1.5%. The authors concluded
that temporal discrimination reflects an aspect of brain func-
tioning that is stronger and more comprehensively related to g
than parameters derived from the Hick RT task.
As already noted, temporal discrimination tasks may invoke
demands on executive cognitive functions. To the extent that this
is so, one would expect a Gt factor to relate strongly with g—and
to a greater degree than RT tasks—because it would be satu-
rated with variance reflecting cognitive functions underpinning
both Gt and g. We have argued that the processes underpinning
performance on temporal discrimination tasks might best align
with memory functions, and previous research has established a
strong and consistent relationship between WM and reasoning
ability; as measured by intelligence tests (e.g. Burns, Nettel-
beck, & McPherson, 2009; Kyllonen & Christal, 1990). Thus,
tasks relying on memory functions should relate strongly to
measures of intelligence, and temporal discrimination tasks
may be an example of such tasks. Put more concisely, memory
functions rather than temporal resolution of the CNS may be
responsible for the relationship between temporal discrimina-
tion and intelligence. We present a reanalysis of Rammsayer
and Brandler’s (2007) data with the aim being to test whether
memory mediates the relationship between temporal discrimi-
nation and intelligence.
Listed in Table 5 are the cognitive ability measures and
temporal discrimination tasks used by Rammsayer and Brandler
(2007) which are relevant to our aims. Participants in their
study were 40 male and 60 female volunteers ranging in age
from 18 to 45 years (M and SD of age: 26.0 ± 6.8 years). The
cognitive measures are composed of subtests of the Leistung-
sprüfsystem (Horn, 1983), Berliner Intelligenztruktur-Test
(Jäger, Süβ, & Beauducel, 1997), and the German adaptation of
Cattell’s Culture Free Test Scale 3 (CCFT; Cattell, 1961; Weiss,
1971). Three of these subtests measure memory functions
(Verbal, Numerical and Spatial Memory). The temporal tasks
and their requirements are described briefly above and more
detailed explanations can be found in the original publication.
A data file containing the correlations, means and standard
deviations reported in Rammsayer and Brandler (2007) was
created for analysis using MPlus 5.2 (Muthen & Muthen, 1998).
I. T. ZAJAC ET AL. 749
Table 5.
Intelligence scales and discrimination tasks used in Rammsayer and
Brandler (2007) a nd the broad ability constructs measured.
Intelligence Tests Broad
Discrimination Tests
Comprehension (VC) Gc Duration
Discrimination (DD1) Gt
Word Fluency (WF) Gc Duration
Discrimination (DD2) Gt
Perceptual Speed (PS) Gs Duration
Discrimination (DD3) Gt
Number 1 (N1) Gs Temporal
Generalisation (TG1) Gt
Number 2 (N2) Gs Temporal
Generalisation (TG2) Gt
Space 1 (SP1) Gf Rhythm
Perception (RP) Gt
Space 2 (SP2) Gf Tonal-order
Judgment (TOJ) Gt
Flexibility of
Closure (CLO) Gf Auditory Flutter
Fusion (AFF)a Gt
Series (SE) Gf
Classifications (CL) Gf
Matrices (MA) Gf
Topologies (TO) Gf
Verbal Memory (VM) Gm
Memory (NM) Gm
Spatial Memory (SM) Gm
Note: Gc, Crystallised Intelligence; Gs, General Speed of Processing; Gf, Fluid
Intelligence; Gm, General Memory; Gt, General Temporal Discrimination, aAFF
excluded from study two analyses.
Confirmatory Factor Analysis (CFA) was then undertaken on
the covariance matrix using Maximum Likelihood estimation.
By using this approach, different models were able to be tested
which either included or omitted a relationship between tempo-
ral discrimination and memory functions, and these were com-
pared using the model chi-square difference test. The fit of
CFA models was assessed using the chi-squared test of model
fit (χ²), the comparative fit index (CFI), the root mean squared
error of approximation (RMSEA), and the standardized root
mean squared residual (SRMR).
We attempted to confirm the presence of the general timing
(Gt) factor reported in Rammsayer and Brandler (2007). How-
ever, we excluded the Auditory Flutter Fusion (AFF) task from
our analysis because it has typically loaded poorly on Gt and
might better be considered a sensory rather than temporal
measure. CFA results indicate that the temporal tasks defined a
Gt factor adequately [χ²(14) = 20.26, p = .122; CFI = .965;
RMSEA = .067; SRMR = .047]. Modification indices sug-
gested that the residuals of TG1 and TG2 should be allowed to
co-vary. Therefore, in an additional model we added this path
and it resulted in a significant improvement in fit [Δχ²(1) = 9.33,
p = .002]. Rhythm Perception (RP) had a weak but significant
loading (r = .37, p < .001) whilst the remaining tasks loaded
strongly with an average of r = .64 (Min = .50, Max = .75, SDr
= .09).
Next, we confirmed the presence of a memory factor by
specifying the three memory tasks to define a single latent Gm
factor. Fit statistics are not available for this model because
degrees of freedom are equal to zero. However, all three tasks
loaded moderately supporting the presence of latent Gm: VM (r
= .56), NM (r = .48), and SM (r = .45).
Rammsayer and Brandler extracted a single psychometric g
(G) factor from the cognitive measures in their study. Instead,
we used a hierarchical model in which specific lower order
factors were defined (Gc, Gs, and Gf; see Table 5) as well as g.
First we attempted to define the lower order factors but statis-
tics showed the model’s fit was not adequate [χ²(51) = 107.83,
p < .001, CFI = .885; RMSEA = .106; SRMR = .072]. There-
fore, in consultation with modification indices, we correlated
the residuals of the Series and Matrices tests. This resulted in a
significant improvement in fit [Δχ²(1) = 32.53, p <. 001], which
was now considered adequate [χ²(50) = 75.29, p = .01, CFI
= .949; RMSEA = .071; SRMR = .059]. The average loading of
the tasks across all factors in this improved solution was r = .69
(min = .51, max = .86, SDr = .11) and the correlations between
the three first order factors were strong. In a subsequent model
the first order factors were used to defined a g factor and the fit
of this hierarchical model was also adequate [χ²(50) = 75.29, p
= .01, CFI = .949; RMSEA = .071; SRMR = .059]. The loading
of each factor on g was strong: Gf (r = .84), Gc (r = .78), Gs (r
= .94).
Having confirmed the presence of temporal, memory and
psychometric factors, we were able to address the extent to
which Gt and Gm are related and predict variance in psycho-
metric g. To accomplish this, we first ran an unrelated predictor
model in which g was regressed onto the independent factors Gt
and Gm. In this model, both Gt (r = .62) and Gm (r = .55) pre-
dicted a significant but comparable amount of variance in g.
Model statistics showed that the fit was not quite adequate
[χ²(202) = 262.99, p < .001, CFI = .921, RMSEA = .055,
SRMR = .086]. Therefore, we tested a related predictor model
in which Gm was regressed onto Gt; whilst still maintaining
regression paths from each of these to g. This related predictor
model resulted in a significant improvement in fit [∆χ²(1) =
7.24, p = .01; χ²(201) = 255.74, p = .005, CFI = .930, RMSEA
= .052, SRMR = .069].
This hierarchical g with related predictors model is presented
as Figure 2. The relationship between Gt and Gm is moderately
strong, with the latent variables sharing approximately 20% of
their variance. This path was necessary for satisfactory fit and
its addition resulted in a marked decrease in the size of the co-
efficient between Gt and g (.47 compared to .62), but not be-
tween Gm and g (.53 versus .55). The standardized direct effect
of Gt on g is .47 and the indirect effect is .23 (.425*.531). Thus,
34% of the effect of Gt on G appears to reflect memory cap-
tured in latent Gm.
In light of the smaller yet significant path between Gt and g
in this related predictor model, we defined a model which ex-
cluded g and instead regressed each of the lower order factors
onto related Gt and Gm factors. The purpose of this was to
better understand the moderate relationship between Gt and g
after accounting for Gm functions. The fit of this model was
good [χ²(197) = 244.91, p = .01, CFI = .938, RMSEA = .049,
SRMR = .067]. The path from Gm to Gc was significant (r
= .57, p <. 001) but not from Gt to Gc (r = .22, p = .11). The
path from Gm to Gs was not significant (r = .19, p = .24) but it
was from Gt to Gs (r = .55, p < .001). Gf was predicted signify-
Figure 2.
Hierarchical g model with related Gt and Gm predictors and standard-
ized parameter estimates.
cantly by Gm (r = .56, p < .001) and to a weaker degree by Gt
(r = .39, p = .003). In a subsequent model we dropped these
non-sig- nificant paths, as well as the covariance between Gs
and Gc because of their relative independence. The fit of this
model decreased significantly [∆χ²(3) = 22.34, p < .001]. How-
ever, overall model fit remained statistically adequate [χ²(200)
= 267.25, p < .001, CFI = .913, RMSEA = .058, SRMR = .078]
and in the interest of parsimony, this more restrictive model—
shown in Figure 2—should be favoured over the former. As can
be seen, Gt relates strongly to Gm (R² = .37) and Gs (R² = .48).
The relationship between Gt and Gf is markedly weaker and
these constructs share only 9% of variance. Gm on the other
hand, relates strongly to both Gc (R² = .67) and Gf (R² = .40).
Discussion—Study 2
The reanalysis in Study 2 has provided a more rigorous as-
sessment of the hypothesis that temporal discrimination tasks
reflect memory functions than Study 1 because of a larger test
battery of cognitive measures. The CFA models show that the
relation of latent Gt to Gm must be incorporated into these
structural models to achieve adequate fit. Moreover, it appears
that around 35% of the relationship between Gt and general
intelligence estimates (g) can be explained by memory func-
tions shared with Gm. Of the three broad cognitive factors ex-
tracted in the second model (see Figure 3), Gt appeared to relate
more strongly to speed of processing (Gs) and Gm than either
Gf, or Gc.
General Conclusion
Recent research has proposed that temporal resolution of the
CNS is partly responsible for intelligent functioning and that
auditory temporal discrimination tasks provide a valid measure
of this resolution (Helmbold & Rammsayer, 2006; Helmbold,
Troche, & Rammsayer, 2006b; Helmbold et al., 2007b;
Rammsayer & Brandler, 2002, 2004, 2007). This paper has
questioned this notion and has presented the results of two
studies designed to elicit the nature of the functions underpin-
ning performance on temporal discrimination tasks.
Study 1 showed that the construct measured by the auditory
TG task is not modality specific. The correlation between visual
and auditory TG was strong; the tasks shared 42% of variance.
Strong relationships have generally not been evident when
adapting ECTs across modalities. For example, the relationship
between visual IT and auditory IT seldom exceeds r = .30 (see
Figure 3.
Broad ability factors model with related Gt and Gm predictors and
standardized parameter estimates.
e.g. Deary, 2000) and the variance in these tasks has been
largely attributed to peripheral sensory type processes (Burns,
Nettelbeck, McPherson, & Stankov, 2007; Burns, Nettelbeck,
& White, 1998; White, 1996; Zajac & Burns, 2007). Contrary
to this, the strong correlation between visual and auditory TG
suggests the processing required by these tasks might not be
sensory but rather cognitively based. This would explain the
relative independence of sensory and temporal discrimination
factors reported previously (Rammsayer & Brandler, 2004).
Study 1 measured the distinct constructs, Gs and WM, to
better understand the observed correlation between TG and
intelligence. Speed of processing was measured using tradi-
tional speed tasks as well as widely researched ECTs (RT and
IT). The impetus for including RT and IT was the proposition
that auditory temporal tasks might be analogous to ECTs in
terms of providing an estimate of neural efficiency (Helmbold
et al., 2007b). The present study does not support this hypothe-
sis. Neither the visual nor auditory TG tasks predicted a statis-
tically significant amount of variance in the composite speed
measure with the shared variance near zero.
The regressions of DM—a measure of working memory –
onto visual and auditory TG tasks were not statistically signifi-
cant. However, given the near equivalence of the TG tasks
across experimental conditions and the consistency of the effect
size, the samples were combined across conditions and visual
and auditory TG did predict a significant amount of variance in
DM; the size of the effect again remained consistent (about
10% shared variance). This shows that the absence of a signifi-
cant effect within experimental conditions reflects a lack of
statistical power. Future studies should increase sample size to
overcome this issue.
The reanalysis of Rammsayer and Brandler’s (2007) data in
Study 2 provide further evidence that temporal discrimination
tasks rely, at least to some extent, on memory functions. In the
hierarchical g model (Figure 2), the path between latent Gt and
Gm factors was both necessary and significant, with the latent
factors sharing 20% of their variance. Furthermore, it was
found that around 35% of Gt’s relationship to g could be attrib-
uted to memory functions represented by latent Gm. In the
second model (Figure 3), the regression of Gm on Gt was
stronger, with the constructs sharing around 37% of their vari-
ance. Of the three broad cognitive ability factors defined, Gt
was most strongly related to Gs. This finding is somewhat con-
sistent with earlier studies in which Gt has been found to share
I. T. ZAJAC ET AL. 751
variance with RT factors. The analysis in earlier studies, how-
ever, has been framed to explore which of Gt and RT explains
more g variance. Not surprisingly, Gt emerges as the stronger
predictor and it almost wholly accounts for the relationship
between RT and g (Helmbold & Rammsayer, 2006; Helmbold
et al., 2007b; Rammsayer & Brandler, 2007).
In the hierarchical model (Figure 2) in Study 2 it appeared
that Gt measured functions over and above memory, which
predicted g variance. The non-hierarchical model (Figure 3)
shows, however, that this significant Gt × g path essentially
reflects Gt’s relationship to Gs, and it is likely because of this
relationship that Gt can account for the correlation between RT
and g. This finding does not imply that Gt measures anything
more fundamental to intelligence than RT tasks; it simply sug-
gests that they measure the same functions.
The relationship reported herein between Gt and memory
measures are consistent with the requirements of the temporal
tasks. Duration Discrimination (DD) requires internal represen-
tations of timed intervals to remain accessible for several sec-
onds following their presentation; TG requires accurate learn-
ing—and thus memorizing—of a standard stimulus, as well as
accurate retrieval of the learnt standard and comparisons with
trial stimuli. Interestingly, the pitch and line-length discrimina-
tion tasks in Study 1 had the same requirements as TG but did
not contribute substantially to the prediction of the working
memory task. One explanation for this finding is that functions
involved in TG are more complex than for pitch and line-length
tasks. Auditory sensory memory, for instance, can retain infor-
mation concerning dimensions like intensity and frequency for
four-to-ten seconds (Jaaskelainen, Hautamake, Naatanen, &
Ilmoniemi, 1999). In the pitch task, then, it is plausible that the
comparisons of stimuli rely heavily on these sensory memory
traces. Conversely, time is not a perceptual dimension but a
cognitively derived entity (Michon, 1990) and therefore the
comparisons of stimuli in auditory and visual TG tasks rely on
cognitive representations of the durations which appear to be
memorised and rehearsed.
Previous research supports this hypothesis. It has been found
in a number of studies that temporal processing of durations
involves the prefrontal cortex (Elbert, Ulrich, Rockstroth, &
Lutzenberger, 1991; Harrington, Haaland, & Knight, 1998)
which is the brain region thought to play a critical role in the
distributed neural systems which achieve working memory
(Engle, Kane, & Tuholski, 1999; Gibbons et al., 2002). Fur-
thermore, an event-related potential (ERP) study which com-
pared temporal and pitch discrimination tasks showed enhanced
prefrontal activation in the temporal task (Gibbons et al., 2002).
This finding was interpreted as indicating a much stronger con-
tribution of executive memory functions to temporal as opposed
to pitch discrimination and it was concluded that “to perceive
time and to evaluate temporal properties of a given stimulus,
formation of cognitive temporal representations is required—a
process primarily based on executive working memory func-
tions” (Gibbons et al., 2002: p. 963).
In summary, the findings herein and those of previous stud-
ies raise questions regarding the extent to which auditory tem-
poral discrimination tasks should be considered measures of
neural efficiency and by extension, temporal resolution of the
CNS. It appears that the observed relationship between auditory
temporal discrimination tasks and measures of g may be ex-
plained almost entirely in terms of memory functions and speed
of processing. More specifically, temporal discrimination per-
formance is confounded by both memory and speed functions
and its relationship to the latter does not automatically imply
that temporal resolution of the CNS is involved. Even if tem-
poral resolution of the CNS is independent of higher order cog-
nitive operations (Rammsayer & Brandler, 2002: p. 509) tempo-
ral discrimination tasks are not. Attempts to gauge the strength
of the relationship between CNS resolution and intelligence—if
indeed there is one—using such tasks will not be unequivocal.
Breitmeyer, B. G. (2007). Visual masking: Past accomplishments,
present status, future developments. Advances in Cognitive Psychol-
ogy, 3, 9-20. doi:10.2478/v10053-008-0010-7
Burns, N. R., & Nettelbeck, T. (2003). Inspection time in the structure
of cognitive abilities: Where does IT fit? Intelligence, 31, 237-255.
Burns, N. R., Nettelbeck, T., & Cooper, C. (2000). Event related poten-
tial correlates of some human cognitive ability constructs. Personal-
ity and Individual Differences, 29, 157-168.
Burns, N. R., Nettelbeck, T., & McPherson, J. (2009). Attention and
intelligence: A factor analytic study. Journal of Individual Differ-
ences, 30, 44-57. doi:10.1027/1614-0001.30.1.44
Burns, N. R., Nettelbeck, T., McPherson, J., & Stankov, L. (2007).
Perceptual learing on inspection time and motion perception. Journal
of General Psychology, 134, 83-100.
Burns, N. R., Nettelbeck, T., & White, M. (1998). Testing the interpre-
tation of inspection time as a measure of speed of sensory processing.
Personality and Individual Differences, 24, 25-39.
Carroll, J. B. (1993). Human cognitive abilities: A survey of factor-
analytic studies. Cambridge, New York: Cambridge University Press.
Cattell, R. B. (1961). Culture free intelligence test, scale 3. Champaign,
IL: Institute for Personality and Ability Testing.
Deary, I. J. (2000). Looking down on human intelligence; from psy-
chometrics to the brain. New York: Oxford University Press.
Elbert, T., Ulrich, R., Rockstroth, B., & Lutzenberger, W. (1991). The
processing of temporal intervals reflected by CNV-loke brain inter-
vals. Psychophysiology, 28, 648-655.
Engle, R. W., Kane, M. J., & Tuholski, S. W. (1999). Individual dif-
ferences in working memory capacity and what they tell us about
controlled attention, general fluid intelligence, and functions in the
prefrontal cortex. In A. Miyake, & P. Shah (Eds.), Models of working
memory: Mechanisms of active maintenance and executive control
(pp. 102-134). Cambridge: Cambridge University Press.
Evans, G., & Nettelbeck, T. (1993). Inspection time: A flash mask to
reduce apparent movement effects. Personality and Individual Dif-
ferences, 15, 91-94. doi:10.1016/0191-8869(93)90045-5
Gibbons, H., Brandler, S., & Rammsayer, T. H. (2002). Dissociating
aspects of temporal and frequency processing: A functional ERP in
humans. Cortex, 39, 947-965. doi:10.1016/S0010-9452(08)70872-4
Grudnik, J. L., & Kranzler, J. H. (2001). Meta-analysis of the relation-
ship between inspection intelligence and inspection time. Intelligence,
29, 523-535. doi:10.1016/S0160-2896(01)00078-2
Harrington, D. L., Haaland, K. Y., & Knight, R. T. (1998). Cortical
networks underlying mechanisms of time perception. Journal of
Neuroscience, 18, 1085-1095.
Helmbold, N., & Rammsayer, T. (2006). Timing performance as a
predictor of psychometric intelligence as measured by speed and
power tests. Jou rnal of Individual Diffe rences, 27, 20-37.
Helmbold, N., & Rammsayer, T. H. (2006). Timing performance as a
predictor of psychometric intelligence as measured by speed and
power tests. Jou rnal of Individual Diffe rences, 27, 20-37.
Helmbold, N., Troche, S., & Rammsayer, T. (2006a). Temporal infor-
mation processing and pitch discrimination as predictors of general
intelligence. Canadian Journal of Experimental Psychology, 60,
294-306. doi:10.1037/cjep2006027
Helmbold, N., Troche, S., & Rammsayer, T. (2007a). Processing of
temporal and nontemporal information as predictors of psychometric
intelligence: A structural-equation-modeling approach. Journal of
Personality, 75, 985-1005. doi:10.1111/j.1467-6494.2007.00463.x
Helmbold, N., Troche, S. J., & Rammsayer, T. H. (2006b). Temporal
information processing and pitch discrimination as predictors of
general intelligence. Canadian Journal of Experimental Psychology,
60, 294-306. doi:10.1037/cjep2006027
Helmbold, N., Troche, S. J., & Rammsayer, T. H. (2007b). Processing
of temporal and nontemporal information as predictors of psycho-
metric intelligence: A structural-equation-modeling approach. Jour-
nal of Personality, 75, 985-1005.
Hendrickson, A. E. (1982). The biological basis of intelligence part I:
Theory. In H. J. Eysenck (Ed.), A model for intelligence (pp.
151-196). Berlin: Springer-Verlag.
Hendrickson, D. E. (1982). The biological basis of intelligence part II:
Measurement. In H. J. Eysenck (Ed.), A model for intelligence (pp.
197-230). Berlin: Springer-Verlag.
Hick, W. E. (1952). On the rate of gain of information. Quarterly
Journal of Experimental Ps ych olo gy, 4, 11-26.
Horn, W. (1983). Leistungsprufsystem. Göttingen, Germany: Hogrefe.
Jaaskelainen, I. P., Hautamake, M., Naatanen, R., & Ilmoniemi, R. J.
(1999). Temporal span of human echoic memory and mismatch
negativity: Revisited. Neuroreport, 10, 1305-1308.
Jäger, A. O., Süβ, H.-M., & Beauducel, A. (1997). Berliner intelligen-
zstruktur test form 4. Göttingen, Germany: Hogrefe.
Jensen, A. R. (1982). Reaction time and psycyometric g. In H. J. Ey-
senck (Ed.), A model for intelligence (pp. 93-132). Berlin: Springer.
Jensen, A. R. (2005). Mental chronometry and the unification of dif-
ferential psychology. In R. J. Sternberg & J. E. Pretz (Eds.), Cogni-
tion and Intelligence: Identifying the mechanisms of the mind (pp.
26-50). Cambridge: Cambridge University Press.
Jensen, A. R. (2006). Clocking the mind: Mental chronometry and
individual differences. Amsterdam: Elsevier.
Kyllonen, P. C., & Christal, R. E. (1990). Reasoning ability is (little
more than) working-memory capacity?! Intelligence, 14, 389-433.
McPherson, J., & Burns, N. R. (2005). A speeded coding task using a
computer-based mouse response. Behavior Research Methods, 37,
538-544. doi:10.3758/BF03192725
Michon, J. A. (1990). Implicit and explicit representations of time. In R.
A. Block (Ed.), Cognitive models of psychological time (pp. 37-58).
Hillsdale, NJ: Lawrence Erlbaum Associates.
Muthen, L. K., & Muthen, B. O. (1998). Mplus. The comprehensive
modelling program for applied research. Users guide. Los Angeles:
Muthen and Muthen.
Nettelbeck, T. (1987). Inspection time and intelligence. In P. A. Vernon
(Ed.), Speed of information processing and intelligence (pp.
295-346). Norwood, NJ: Ablex.
Nettelbeck, T. (2001). Correlation between inspection time and psy-
chometric abilities. A personal interpretation. Intelligence, 29, 459-474.
Nettelbeck, T. (2003). Inspection time and g. In H. Nyborg (Ed.), The
scientific study of general intelligence: Tribute to Arthur R. Jensen
(pp. 77-92). Amsterdam, Boston: Permagon.
Neubauer, A. C., & Fink, A. (2005). Basic information processing and
the psychophysiology of intelligence. In R. J. Sternberg, & J. E. Pretz
(Eds.), Cognition and intelligence: Identifying the mechanisms of the
mind (pp. 68-87). Cambridge: Cambridge University Press.
O’Connor, T. A., & Burns, N. R. (2003). Inspection time and general
speed of processing. Personality and individual differences, 35,
713-724. doi:10.1016/S0191-8869(02)00264-7
Petrill, S. A., Luo, D., Thompson, L. A., & Detterman, D. K. (2001).
Inspection time and the relationship among elementary cognitive
tasks, general intelligence, and specific cognitive abilities. Intelli-
gence, 29, 487-496. doi:10.1016/S0160-2896(01)00074-5
Rammsayer, T. H., & Brandler, S. (2002). On the relationship between
general fluid intelligence and psychophysical indicators of temporal
resolution in the brain. Journal of Research in Personality, 36,
507-530. doi:10.1016/S0092-6566(02)00006-5
Rammsayer, T. H., & Brandler, S. (2004). Aspects of temporal infor-
mation processing: A dimensional analysis. Psychological Research,
69, 115-123. doi:10.1007/s00426-003-0164-3
Rammsayer, T. H., & Brandler, S. (2007). Performance on temporal
information processing as an index of general intelligence. Intelli-
gence, 35, 123-139. doi:10.1016/j.intell.2006.04.007
Reed, T., & Jensen, A. (1993). Choice reaction time and visual pathway
nerve conduction velocity both correlate with intelligence but appear
not to correlate with each other: Implications for information proc-
essing. Intelligence, 17, 191-203.
Sheppard, L. D., & Vernon, P. A. (2008). Intelligence and speed of
information-processing: A review of 50 years of research. Intelli-
gence, 44, 535-551.
Stankov, L. (2005). Reductionism versus charting. In R. J. Sternberg &
J. E. Pretz (Eds.), Cognition and intelligence: Identifying the mecha-
nisms of the mind (pp. 51-67). Cambridge: Cambridge University
Surwillo, W. W. (1968). Timing of behaviour in senescence and the
role of the central nervous system. In G. A. Talland (Ed.), Human
aging and behavior (pp. 1-35). New York: Academic Press.
Thurstone, L. L. (1938). Primary mental abilities. Chicago, IL: Univer-
sity of Chicago Press.
Vernon, P. A. (1993). Intelligence and neural efficiency. In D. K. Det-
terman (Ed.), Current topics in human intelligence (Vol. 3, pp.
171-188). New Jersey: Ablex Publishing Corporation.
Weiss, R. H. (1971). Grundintelligenztest CFT 3 skala 3. Braunschweig,
Germany: Westermann.
Wetherill, G. B., & Levitt, H. (1965). Sequential estimation of points
on a psychometric function. British Journal of Mathematical and
Statistical Psychology, 18, 1-10.
White, M. (1996). Interpreting inspection time as a measure of the
speed of sensory processing. Personality and Individual Differences,
20, 351-363. doi:10.1016/0191-8869(95)00171-9
Zajac, I. T., & Burns, N. R. (2007). Measuring auditory inspection time
in primary school children. Journal of Individual Differences, 28,
45-52. doi:10.1027/1614-0001.28.1.45
I. T. ZAJAC ET AL. 753
Appendix A
Frequency distribution of stimuli in the stimuli set presented within the Dissociation Paradigm.
Dimension 1 Stimulus Duration
Dimension 2 125 ms 150 ms 175 ms 200 ms (S) 225 ms 250 ms 275 ms
964 Hz 1 1 1 3 1 1 1 9
976 Hz 1 1 1 3 1 1 1 9
988 Hz 1 1 1 3 1 1 1 9
1000 Hz (S) 3 3 3 9 3 3 3 27
1012 Hz 1 1 1 3 1 1 1 9
1024 Hz 1 1 1 3 1 1 1 9
1036 Hz 1 1 1 3 1 1 1 9
Σ 9 9 9 27 9 9 9 81
7 cm 1 1 1 3 1 1 1 9
8 cm 1 1 1 3 1 1 1 9
9 cm 1 1 1 3 1 1 1 9
10 cm (S) 3 3 3 9 3 3 3 27
11 cm 1 1 1 3 1 1 1 9
12 cm 1 1 1 3 1 1 1 9
13 cm 1 1 1 3 1 1 1 9
(Line Length)
Σ 9 9 9 27 9 9 9 81
Note: S = Standard.