Reactions and Reaction time overview

This overview describes two primary types of human reactions pertinent to the case; involuntary reactions and voluntary reactions.

Involuntary actions/reactions and the startle reflex

Summary points from this startle reaction discussion:


An involuntary action is one which occurs without the conscious choice of an organism. If it occurs specifically in response to a stimulus, it will be known as a reflex. The reflex example of interest here is the startle reflex as a reaction to loud auditory stimuli.

In 1929 Strauss, using a revolver shot as a stimulus, isolated a definite startle behavior pattern. It was definite, symmetrical, and relatively uninfluenced by postural changes. One of its most striking characteristics was its rapidity: it might come and go within half a second.

In 1939 Landis and Hunt, using pistol shots and high speed photography, conducted similar tests which confirmed the Strauss results. The reaction pattern included blinking of the eyes, head movement forward, a characteristic facial expression, raising and drawing forward of the shoulders, abduction of the upper arms, bending of the elbows, pronation of the lower arms, flexion of the fingers, forward movement of the trunk, contraction of the abdomen, and bending of the knees. Likewise they saw the typical response happen in about a half a second (but it could range from 0.3 to over one second depending on individuals and intensity). Landis and Hunt called the startle pattern an immediate, involuntary, and general flexion behavior.

A.N. Carlsen described the startle reflex in a thesis of “Auditory Startle Response and Reaction Time” as a generalized and diffuse protective response consisting of a characteristic set of muscle actions initiated by a sudden intense stimulus. More specifically, in response to a sudden, unexpected acoustic, tactile or vestibular stimulus, a generalized flexion response is observed in mammals. The startle response consists of a characteristic pattern of muscle flexion, as well as an increase in central nervous system and autonomic activity.

Acoustic stimuli must be adequately loud (at least 90 dB) to elicit a startle response although more intense stimuli produce larger amplitude responses and shorter response latencies.

Landis and Hunt described the startle response as a patterned response consisting of several bilateral stereotyped muscle movements. This response started with blinking of the eyes and a characteristic facial expression, along with dorsiflexion of the head and neck. The described response included a curling of the shoulders in a ventro-caudal direction, flexion of the elbows and fingers, bending of the trunk and bending of the knees. This generalized flexion response has been hypothesized by Yeomans and Frankland to be an adaptive defense response in terrestrial mammals to predatory attack from the rear, as the response results in reduced exposure of the dorsal surface of the neck, a vulnerable point of attack. Brown et al. described a similar response pattern consisting of eye closure, grimacing, as well as neck, trunk and limb flexion.

Important final points regarding startle reaction.
The definitions below are useful in understanding body movements related to joint articulation:
Flexion: Bending the joint resulting in a decrease of angle; moving the head forward at the joint just below the skull.
Rotation: Rotary movement around the longitudinal axis of the bone; turning the head to the side (right or left) at a joint below the skull.
In every study reviewed here on startle reaction,  all the startle movements are classified as flexion movements. This translates to the head moving down and forward (which effectively maintains symmetry in the front/back plane going through the spine or sagittal plane). There are no studies indicating rotation (right or left) as a startle response.


Key conclusion: None of the rapid head rotations seen in the Zapruder film are startle reactions, but rather voluntary or secondary reactions of rotation.


Voluntary actions/reactions

Summary points from this voluntary reaction discussion:


Voluntary action is an anticipated, cognitive goal-orientated movement. Voluntary action is demonstrated when one cognitively identifies the desired outcome and pairs it with the action it will take to achieve it.

Voluntary reaction time can be based and tested on reacting to simple or complex stimulus scenarios.  Simple voluntary reaction time would include reacting to one stimulus and having one response (1). An example of this could be the reaction to sound with a head, hand, or foot movement. This is the simplest kind of reaction time, with little or no decision making (3) and is very relevant to the conditions encountered in the motorcade with rifle reports and subsequent reactions.

Voluntary response time, sometimes called Perception-Response Time or Perception-Reaction Time (PRT), is the lag in time between detection of an input (stimulus) and the start of initiation of a controlled response. Response time and Reaction time are sometimes used interchangeably, but typically the reaction time is the time to initiate reaction movement and response time is reaction time plus an additional short period of time for the reaction movement time (3).
Human voluntary response time is a complex dynamic and can be dependent on a number of variables. Robert Kosinski (1) conducted a literature review on reaction time and the studies which looked at factors that influence reaction time. A number of factors have been studied and included variables like muscular exercise/arousal, age, gender, fatigue, alcohol, drugs, Illness, muscle tremors, previous brain injury, fasting, etc.

For the occupants of the Presidential limo and Secret Service follow up car entering Dealey Plaza, there likely wasn’t much variation from normal values for these listed variables, and over the duration of time the three shots were fired, except for JFK and Connally who had injuries, none of these factors changed.  There is however, a variable that has been shown to be a significant factor on reaction time, and would have changed over the timeframe of the assassination. Specifically this variable is “warning time” that a stimulus will occur soon. Kosinski’s references showed that that reaction times are faster when the subject has been warned that a stimulus will arrive soon, and the warning can be fast in advance. This variable has been shown significant relative to reaction time in a number of studies and can be classified as levels of “expectedness”. This variable would be important during the time of Zapruder’s filming and warrants further evaluation as the state of awareness or expectedness of shots or distress in the motorcade changed significantly and quickly as the assassination progressed.

Since expectedness of the stimulus would be the reaction time variable that changed for individuals in the assassination sequence, it would be ideal to find a study that related Perception-Reaction Time (PRT) to Stimulus Expectancy and compare that to the onset of reactions observed in the Zapruder film. Unfortunately no study controlled in exactly this way could be found. This is not a surprise as controlling an experiment with expectancy level as an independent variable may not be easy.

In an effort to extract some data on reaction time to stimulus expectancy, this author conducted a Meta-analysis on literature that had experiments on reaction time to audible stimulus, with an objective to find studies that appear to have differences in the panelist’s degree of expectancy. Reaction time studies were reviewed,  and if relevant, were classified into one of 4 groups of panelist’s stimulus expectancy (Highly Expected, Expected, Unexpected, and Complete Surprise).

These groups were classified as follows:
Highly Expected: Subjects know almost exactly when the stimulus will occur (pre-knowledge on impending stimulus timeframe: typically seconds)
Expected: Subjects know the stimulus may or will occur soon, but don’t know exactly when it will happen (pre-knowledge on impending stimulus timeframe: typically seconds-minutes)
Unexpected: Subjects know a stimulus may or will occur within an extended timeframe, but didn’t know when it would happen (pre-knowledge on impending stimulus timeframe: typically minutes-hours)
Complete Surprise: Subjects not aware a stimulus will occur (pre-knowledge on impending stimulus timeframe: unaware or long enough to essentially be forgotten)

The results of this survey are summarized below with the mean reaction time values from the studies plotted above the approximate expectedness category.

The Meta analysis result shows a range in voluntary reaction time, depending on the degree of pre-awareness, from just over 100 ms to somewhat over 1100 ms.  Perhaps not surprisingly, the greatest change in reaction time appeared to occur around the transition from an Expected to an Unexpected state of awareness of the impending stimulus.

Voluntary reactions can occur in conjunction with startle reactions.  When both type of reactions occur, the voluntary reaction is typically called a secondary reaction as the startle reflex starts first. The voluntary secondary reactions typically begin after startle reactions have subsided, but can begin during the progression of the primary startle reactions.
Landis and Hunt classified the voluntary secondary reactions that follow a startle reaction into four categories. Curiosity, Fear, Annoyance, and Overflow effects.
Curiosity would include not only regard and attention to the stimulus source but also to other elements of the situation and even to the purpose of the experiment;  the behavior would range from turning of the head and body toward the stimulus source to a mere quizzical raising of the eyebrows.  Fear would include actual flight behavior as well as mere defensive gestures, and would cover such behaviors as running away and covering the face or ears with the hands. Annoyance would not reach the severe overt expression of anger such as attack, but there are subjects who through gestures, speech, or facial expression indicate their displeasure and irritation. The final category of Overflow effect would include all those cases where the behavior does not seem to be rational.

Both head and body turning is seen in the early frames of the Zapruder film (see reference page on early film reactions). Additionally, displeasure and irritation may be apparent (see John Connally and Jackie Kennedy facial expressions after their heads turn left at about z161, best seen in the Croft close up photo).



Reaction Time Survey: Mean time to onset of voluntary or secondary reactions for various studies

When the study mean reaction time included a short hand or foot movement to trigger a switch, an estimate of that movement time was subtracted to get an estimate of the onset of the voluntary motion/reaction. For example, if applicable, a typical movement time might be about 0.2 seconds.

Reference Study

Authors; Title

(including the assigned category and mean reaction time)


Brown, Kenwell, Maraj & Collins;
“Go” Signal Intensity Influences the Sprint Start
Medicine & Science in Sports & Exercise 2008

Sprinters and reaction times. Secondary reaction onset is taken as onset of pressure on starting blocks. A Startle reaction may help facilitate a secondary reaction by just a small amount (~10ms) but this is important in Olympic sprint times.
Highly Expected; 128ms


Landis & Hunt
The Startle Pattern 1939

Secondary movement test (small subset of broader study) where subjects told to get ready and “try to jump” when the shot sounded.
Highly Expected / Expected; 152ms


Human Benchmark – Reaction Time

Online test of reaction. When a color changes within a couple of seconds of test start, mouse click as fast as possible.
Highly Expected; 215ms


Human Factors

Summary of PRT from literature review for drivers with a “expected” stimulus. Drivers aware of impending signal to brake but not exactly when.
Taken from 16 studies.
(0.2 seconds for foot movement, based on Marc Green data, subtracted from 540 ms)
Expected; 340ms


Landis & Hunt
The Startle Pattern 1939

Subjects knew a test was near (they were wired up for reflex measurements)
Representative Secondary facial movements noted to begin after startle.
Expected; 305ms


Johansson & Rumar
Drivers Brake Reaction Time 1971

Subjects knew a buzzer in car would sound for a brake movement, but not sure when it would happen (at least an hour).
Unexpected; 530ms


Eckman, Friesen & Simons
Is the Startle Reaction an Emotion? 1985

Subjects knew a pistol would be fired in some timeframe within about an hour, but not when.

Unexpected; 1100ms


Human Factors

Summary of PRT from literature review for drivers with a “surprise” stimulus. Drivers don’t even know when or if a stimulus will happen. Taken from 16 studies.
(0.2 seconds for foot movement, based on Marc Green data, subtracted from 1300 ms)
Surprise; 1100ms



Synopsis taken from Johansson Study reference above (F)

When drivers tested had a high degree of expectancy, the mean brake reaction time was as low as 0.45 seconds. The perception time part was about 0.25 seconds (hence 0.15 to 0.2 seconds for foot movement).
Highly Expected –Expected; 250ms


Warrick, Kibler, & Topmiller
Synopsis taken from Thackray reference below (L)

Secretaries to press a button nearby when buzzer sounds. Un-alerted it took about 0.8 sec. (33% longer than alerted)
(I then subtract 0.2 for hand movement)
Unexpected; 600ms


Warrick, Kibler, & Topmiller
Synopsis taken from Thackray reference below (L)

Secretaries to press a button nearby when buzzer sounds. Alerted it took about 0.6 sec.
(I then subtract 0.2 for hand movement)
Expected; 400ms


Performance Recovery Following Startle: A laboratory Approach to the Study of Behavioral Response to Sudden Aircraft Emergencies 1988

A first loud sound burst had a mean response time of 893 ms. in lab test.
(I then subtract 0.2 for hand movement)
Unexpected-Expected; 693ms



Kosinski A literature review on reaction time