The "overheat" (flushing of the skin) you experience is a common-enough reaction to stress, part of the fight-or-flight response. It's not uncommon for people to experince it during exams; generally exam/test axniety is related to both internal and external factors (or perception thereof)
Self-esteem was a significant and strong predictor of test anxiety. Perceived difficulty of the test and the high-stakes nature or consequences of the test was also related to higher test anxiety.
It's beyond the scope of this site to suggest individualized treatments, but generally you may want to read about anxiety disorders and panic attacks, more specifically.
Although there are vast amounts of academic publications about the measurement and the effects of test anxiety, the amount of publications trying to assess coping strategies appears a lot more limited. E.g. I found a small study about relaxation techniques, Larson and Rose (2011), which found them effective on reducing the anxiety, but ultimately having no impact on test resuts.
Although I usually recommend Wikipedia as a good first read on many topics, their article on test anxiety is pretty bad, especially in the treatment section, with a heavy emphasis on medication... almost entirely sourced from a poor-quality site. That's in sharp contrast for instance with NHS' take on the matter, which makes virtually no mention of medication. Even the (US) Mayo clinic for instance doesn't place much emphasis on medication for test anxiety.
And what I discovered in the last two paragraphs turns out is also reflect in others' conclusions e.g. UCLA's Center for Mental Health in Schools:
In a 2013 review, von der Embse, Barterian, and Segool (2013) concluded that “there are few studies
that have examined test-anxiety interventions with elementary and secondary school students.
However, techniques including biofeedback, behavior therapy, cognitive behavioral therapy, priming
competency, and mixed approaches have demonstrated promising results.” While medication also
is discussed, it is not the preferred treatment and, if prescribed, is to be used in conjunction with
psychological interventions (Tracy, 2012).
Alas this UCLA source/quote is focused on pre-university education and tests. However the paper of von der Embse pointed to an older meta-analysis that didn't have this narrower focus, Eregne (2003), which concluded that
The treatment of test anxiety has been quite successful in reducing the test anxiety level of clients. The most effective treatments appear to be those that combine skill-focussed approaches with behaviour or cognitive approaches. Individually conducted programs, along with programs that combined individual and group counselling formats, produced the greatest changes.
And this older paper (whose primary studies were almost exclusively aimed at university students) has a fairly detailed assessment of the effectiveness of various types of interventions (although I personally wonder if there was enough statistical power for each kind to draw such definitive conclusions):
Test anxiety reduction groups was
categorized into 21 different interventions. The observed differences
among these interventions were significant. Type of the intervention
was related to the effectiveness of the intervention. Behavioural treatments
included systematic desensitization, relaxation training, anxiety
management training, hypnosis and other behavioural treatments such
as modelling and extinction. Systematic desensitization and other
behavioural techniques were found to be effective, E+ = 0.90 and E+ = 1.01
respectively. Relaxation training alone and hypnotherapy were
moderately effective, E+ = 0.50 and E+ = 0.52 respectively. Although
Hembree’s (1988) meta-analytic study supported the effectiveness of
relaxation training, this finding has been contradicted in the literature.
Among cognitive therapies, cognitive restructuring produced the
largest effect size (E+ = 1.11). However, rational-emotive therapy and
other cognitive techniques produced moderate effect sizes, E+ = 0.54 and
E+ = 0.48 respectively. These findings are also consistent with previous
studies (DiTomasso, 1980; O’Bryan, 1986; Thompson, 1987). These
interventions may be effective on the emotionality and ‘worry’ components
of the test anxiety construct as proposed by Spielberger and Vagg (1995).
The result of this meta-analysis indicate that effectiveness of cognitive
therapies increased when they were combined with skill-focussed
techniques. This finding is consistent with O’Bryan’s (1986) metaanalytic
study. Clients may increase their sense of readiness and their
confidence level by learning different study and test taking skills.
Cognitive behavioural interventions were found to be small to moderately
effective in reducing test anxiety. This contradicts the findings of the
studies conducted by DiTomasso (1980) and Gambles (1994).
Study skills training alone had a small effect size (E+ = 0.28). This
indicates that test anxiety cannot be explained by a learning deficit
model which indicates the lack of effective study skills and test taking
skills. While cognitive and behavioural treatment techniques focus on
relieving symptoms of test anxiety that interfere with an individual’s
ability to perform well or be comfortable in testing situations, other
researchers have conceptualized test anxiety as a result of a deficit of
knowledge, training, ability to study, or test-taking skills. These theorists
suggest that improving an individual’s skills in these areas will result in
improved performance and comfort. Results of this meta-analysis indicate
that this may not necessarily be the case. Study skills training when
measured by test anxiety show only small and moderate effect sizes
(E+ = 0.28, E+ = 0.43, E+ = 0.45). These results indicate that subjects
treated with study skills training have only moderate improvement
when compared with individuals who receive no treatment. However, a
trend is indicated for significant improvement when study skills training
is combined with cognitive or behavioural interventions. One of the
significant findings of this study is that effectiveness of behavioural or
cognitive interventions is increased whenever they are combined with
study skills training.
Although a number of interventions have been suggested and tested in
test anxiety reduction, an integrated approach, utilizing various strategies
in combination, is likely have special value. Such a combined program
might include such components as: (1) applicable information about
studying and test taking skills; (2) opportunity to observe someone take
tests, or test-like problems; (3) increasing self-monitoring and practice
self-control especially of behaviour and thought in test-taking situations;
(4) practice in attending to the task in hand, attentional training and
(5) learning how to relax under test specific conditions.
Modality. The test anxiety reduction programs which were done with
groups and individuals together produced larger effect sizes (E+ = 0.84);
second, the group format had an effect size of (E+ = 0.67). The individual
format had the smallest effect size (E+ = 0.34). Most of the intervention
programs (83 percent) were conducted in the group format. It is highly
likely that clients in groups can get other students’ perceptions and get
alternative solutions to their problems. They may see that they are not
the only person who has problems related to test anxiety; they may gain
therapeutic benefit in group formats and easily transfer it into their
daily lives.
Also his ("E+") measure of effect size is based on Hunter and Schmidt (1990) methodology, which itself has been criticized by some for allegedly inflating effect sizes.
