View Full Version : Interference in task learning/specificity of training

Tony Blazevich
10-03-1999, 04:31 PM
Dear members,

Several weeks ago I asked for information regarding task performance after
learning either in isolation or concurrently with other tasks. I had
performed a resistance training study where subjects also performed sprint
running and vertical jump training concurrently. I found no differences
between my groups in either the running/jumping or resistance tasks after a
short period of training. This is despite other studies using only one
form of training reporting significant training-specific adaptations to
resistance training. Now that I have had time to review the literature I
am able to provide a detailed summary. A brief account of my study is
presented below, it may be easier to read the following if you have
reminded yourself of the study. Any discussion or feedback is welcome.

First, it is important to discriminate between two theories of learning.
The first is the specificity of learning principle that states that
performance in a task is better when the training for that task is more
specific or similar. Second is the contextual interference theory that
proposes that learning is enhanced when practice similar, but not
identical, tasks are presented in a random order. From the studies I was
able to find, I noticed that research into the specificity of learning
principle used a methodology where several practice tasks were used in
isolation by different groups. Those who trained with a task very similar
to the test task had the best improvements in performance. However,
research into contextual interference used methodologies where several
tasks were performed concurrently in training. Results suggest that task
learning is enhanced when several different tasks are performed in a random
order than when one task is performed in isolation in training. This
second methodology most closely resembles the methodology used in my
training study. In my study, I had three groups, two of which performed
resistance training and sprint/jump training (although they used different
resistance training exercises) and one that performed only sprint/jump
training. Using the theories of contextual interference, the groups should
have been able to perform all the tests simlarly (including the resistance
exercises since they were chosen to be similar to either the jump or sprint
movements) since all the groups trained using several different tasks.
Their ability to perform at any task that is similar should then have been
improved. It is important to note that the short period of 'specific'
training (5 weeks, they also performed 4 weeks of non-specific training
first) was not sufficient to stimulate gains in muscle size, therefore any
adaptation should have been largely neural, or due to learning. If the
contextual interference between the groups was simliar, then all groups
should have improved the same.

A second piece of information that was important in explaining the outcome
of my research was provided by Daniel Ferris. He emailed me two abstracts.
The first was by Brashers-Krug et al. (see below for reference). The
authors found that the learning of a second motor task was not disrupted if
learned 4 hours after another since the delay allowed for consolidation.
Also, Norrie and Henry (see below for reference) found that if a novel task
was performed during rest periods in the learning of a task, performance in
the learned task was significantly reduced. This is possibly because
consolidation did not occur during rest. In my study, the sprint/jump and
resistance sessions were performed on different days, therefore there was
equal consolidation between the groups. The group that performed no
resistance training performed sprint and jump training only, but both in
the same session. One might have predicted that since this group performed
twice the number of sprint/jump training sessions, they might have improved
more in a short period of time that the resistance groups (expecially since
only small strength improvements in the resistance groups might have been
possible). However, the interference caused by practicing two tasks in
succession with no period of consolidation could have hampered the task
improvements in this group.

A third piece of important information was from a PhD dissertation
performed by Daniel Ferris (dferris@ucla.edu). He found that performance
in a circle-forming motor task was simliar between groups who practiced
with an easy, complcated or several combined methods of drawing the
circles. He concluded that learning occurs at a perceptual rather than
specific level regardless of the details of the practice. This suggests
that in my study, there should be no difference between the resistance
training groups. While the movement patterns of the two exercises
differed, the subjects were able to use the lessons learned in practice for
one task to improve their performance in another. Unfortunately I could
not find any similar research in the SportDiscus or PsychLit databases. I
therefore could not determine why tasks seemed to be learned at a
'perceptual level' when a large body of research suggests that adaptations
to task training can be very specific.

However, it is now possible to speculate as to why I did not find any
difference among my training groups. First it is important to note that
performing concurrent strength and sprint/jump training is very different
from performing such discrete tasks as those used in many motor control
studies. Nonetheless, the theories from motor learning seem to be able to
explain why I found no between-group differences in my study when other
studies that only examined the influence on training of one specific task
showed that the adaptations to that task were very specific. First, all
groups had contextual interference in their training allowing for a similar
improvement in performance between the groups. Second, it was likely that
a similar performance improvement would occur between the resistance groups
because the task and resistance sessions were performed on different days.
This allowed for consolidation. The group that performed twice as many
sprint/jump sessions did not improve more than the resistance groups (which
could be likely given no muscle size increases and thus little actual
muscle force increases were likely, and the resistance groups only
performed two sprint/jump sessions per week) possibly because the
sprint/jump sessions were performed with no rest between the training of
the two tasks. This would have limited the consilidation of each task and
limited improvements. Third, it is possible that learning of complex tasks
occurs at a more perceptual level than previously thought. Therefore
subjects were able to use 'techniques' learned on one task to successfully
improve other similar tasks. This minimised any differences between the

Of course it is also important to mention that the subjects were all good
athletes from a back ground generally of soccer, rugby, rugby league or
netball and all subjects performed 4 weeks of supervised resistance and
sprint/jump training prior to the 5-week specific training phase.
Therefore, the sprint and jump techniques of all subjects would have been
good and unlikely to change significantly in the following 5 weeks of
training. As far as the resistance training is concerned however, research
into movement-pattern specficity of training suggests that even well
trained strength athletes improve rapidly when training at a novel strength
task. Thus I might have expected differences in strength performance
between my resistance groups where each group improved most at the exercise
they used in training. The fact that there were no differences between my
resistance groups might be explained by the three reasons described above.

Authors mentioned in responses:

Shea & Kohl (especially 1990's)
Shea & Wright (especially 1990's)
Susan Higgins
Claude Ghez
Jim Gordon
Timothy Lee
Guadagmoli, M
Reza Shadmehr (John Hopkins University)
Franklin Henry
Gentile, AM (Medicine and Science in Sports, 1992)
Magill & Hall (Human Movement Science, 1990)

Noted references:

Brashers-Krug T, Shadmehr R, Bizzi E. Consolidation in motor memory. Nature
July: 18, 382(6588): 252-255, 1996.

Norrie ML, Henry FM. Influence of an interpolated non-related motor task on
shrot- and long-term memory learning and retention of a gross motor skill.
Perceptual and motor skills June: 46(3 Pt 1): 987-994, 1978.

Overview of training study and results:

Adaptations to resistance training are specific to the velocity and
movement pattern of the training exercises. For example, if someone trains
with a particular exercise (eg a cable tricep extension) the strength
gained does not transfer well to other, similar, exercises (eg an overhead
tricep extension). Furthermore, these 'specific' strength adaptations can
occur over a short period of time (many studies have shown joint
angle-specific adaptations to training after training periods of only 4-6
weeks [Kitai & Sale, 1989; Thepaut-Mathieu et al., 1988; Weir et al., 1994,
1995]). Thus many athletes try to mimic their sporting movements when
performing resistance training. However it has not been shown
experimentally whether 'specific' resistance training is associated with
greater improvements in a sporting task than 'non-specific' resistance
training when both resistance and task training are performed concurrently.
I therefore conducted a 5-week (+ 4-week 'non-specific' lead-in training)
training study to examine short-term changes in sprint and vertical jump
performance after a period of concurrent task and resistance training.
Briefly, three groups performed sprint and vertical jump training with one
of these groups performing moderate to high speed squat lift training and
one of these groups performing a new exercise (coined the 'forward hack
squat': a unilateral squat-like movement performed with the body prone to
45 deg and the hip and knee angles moving through very similar ranges of
motion to those in the acceleration phase (0 - 20 m) of running) in
addition to their running training. Despite fully supervised training and
a high training attendance by subjects we found no difference between the
groups in neither sprint or jump performance, nor in their performance in
the resistance exercises.

Given that the movement patterns of the resistance exercises were so
different, I expected large inter-group differences even if there were no
differences in sprint/vertical jump performance between the groups.
Isokinetic testing showed changes in the angles at which the groups
produced maximum torque in an isokinetic knee extension such that the squat
group acheived their maximum at a more closed angle after training while
the forward hack squat group acheived their maximum at a more open angle.
This difference however only approached significance (p=0.054).
Nonetheless it suggests that some adaptations may have occurred and that
these adaptations differed between the groups. It is my theory that
perhaps performing sprint/jump training concurrently with resistance
training interfered with the learning of the resistance tasks. I would
assume that if the groups only performed weight training we would have seen
differences between the groups in resistance exercies performance. Perhaps
if the training period had been longer we might still have seen
training-specific adaptations, but these adaptations might have occurred
slower during concurrent resistance and sprint/jump training than when only
resistance training is being performed. I therefore would like to access
literature examining the time-course of learning when tasks are practiced
in isolation and concurrently with other tasks. I am using the term
learning because I assume that any change in performance within the first
few weeks would be due to changes within the nervous system affecting
muscle recruitment patterns etc., i.e learning.

Thanks for your help.


************************************************** ****************
Anthony (Tony) Blazevich
School of Exercise Science and Sport Management
Southern Cross University
Lismore NSW Australia, 2480

Email: ablaze20@scu.edu.au
Phone: +61 2 6620 3231
Fax: +61 2 6620 3880
Mobile: 0416 200 674 (Australia only)

************************************************** ****************

To unsubscribe send SIGNOFF BIOMCH-L to LISTSERV@nic.surfnet.nl
For information and archives: http://isb.ri.ccf.org/biomch-l