- Open Access
The influence of emotion on keyboard typing: an experimental study using visual stimuli
© Lee et al.; licensee BioMed Central Ltd. 2014
- Received: 6 May 2014
- Accepted: 11 June 2014
- Published: 20 June 2014
Emotion recognition technology plays the essential role of enhancement in Human-Computer Interaction (HCI). In recent years, a novel approach for emotion recognition has been reported, which is by keystroke dynamics. This approach can be considered to be rather desirable in HCI because the data used is rather non-intrusive and easy to obtain. However, there were only limited investigations about the phenomenon itself in previous studies. This study aims to examine the source of variance in keystroke typing patterns caused by emotions.
A controlled experiment to collect subjects’ keystroke data in different emotional states induced by International Affective Picture System (IAPS) was conducted. Two-way Valence (3) × Arousal (3) ANOVAs were used to examine the collected dataset.
The results of the experiment indicate that the effect of emotion is significant (p < .001) in the keystroke duration, keystroke latency, and accuracy rate of the keyboard typing. However, the size of the emotional effect is small, compare to the individual variability.
Our findings support the conclusion that the keystroke duration, keystroke latency, and also the accuracy rate of typing, are influenced by emotional states. Notably, the finding about the size of effect suggests that the accuracy rate of the emotion recognition could be further improved if personalized models are utilized. On the other hand, the finding also provides an explanation of why real-world applications which authenticate the identity of users by monitoring keystrokes may not be interfered by the emotional states of users. The experiment was conducted using standard instruments and hence is expected to be highly reproducible.
- Keyboard typing
- Human subject experiment
- International affective picture system
Graphics and the computing capabilities of computers have become powerful recently. However, a computer interactive application that does not understand or adapt to a users’ context, such as the emotion states of a user, could still lead to usability problems. Such an application could provide annoying feedback, interrupt users in an inappropriate situation, or increase the user’s frustration. Furthermore, it is also known that emotion can affect people with respect to their memory, assessment, judgment, expectations, opinions and even motor behaviors. Hence, it is crucial to consider the effect of emotions in modern usability studies: a computer could take advantage of the effect by presenting stimuli that sustain the desired emotions or, alternatively, avoid undesired emotions. Frustrated users could be guided to a different task, focus on a different aspect of the current task, or simply be advised to take a break.
In 1990s, Rosalind W. Picard, the mother of “Affective Computing”, began to propose and demonstrate her ideas about having computers identify a user’s emotion states and about possible improvements to the usability of computer applications . Subsequently, many approaches for detecting users’ emotions have been demonstrated to be useful. For example, emotion recognition by facial expression, which aims to model visually distinguishable facial movements ; by speech, for which researchers utilize acoustic features such as pitch, intensity, duration, and spectral data ; and by physiological data, such as the heart rate and sweat .
Emotion recognition technology based on keystroke dynamics was not reported in the literature until Zimmermann P, Guttormsen S, Danuser B and Gomez P  first described this approach. These authors proposed an experiment designed to examine the effect of film-induced emotional states (PVHA, PVLA, NVHA, NVLA and nVnA (P = positive, N = negative, H = high, L = low, n = neutral, V = valence, A = arousal)) in subjects, with the keystroke dynamics in regard to keystroke rate per second, average duration of keystroke (from key-down until key-up event). However, they did not actually carry out the work described in their proposal. The use of keystroke dynamics for emotion recognition has two main advantages that make such the technique favorable. The two advantages are that it is non-intrusive and easy-to-obtain because the technique does not require any additional equipment or sensors other than a standard input device, which is the keyboard of a computer.
Later, numerous studies in the field of computer science have reported the development of emotion recognition technology based on keystroke dynamics. Vizer LM, Zhou L and Sears A  reported the use of ratios between specific keys and all keys to recognize task-induced cognitive and physical stresses from a neutral state. They achieved a classification rate of 62.5% for physical stress and 75% for cognitive stress. The key ratios could represent the frequencies of typing specific keys, which may increase or decrease due to the changes in emotional state. The analysis result was produced based on sophisticated Machine-Learning (ML) algorithms, and hence, the relationship between emotion and these ratios was not identified. Notably, most of the main streams of ML algorithms only produce models that are considered to be a black box, and do not produce readablea models. The ML algorithms are usually used for building models from dataset that contains complex relationships which are not able to be identified by a traditional statistical model (e.g., t-test, ANOVA).
In 2011, Epp C, Lippold M and Mandryk RL  reported a result of building models to recognize experience-sampled emotional states based on keystroke durations and latencies that were extracted from a fixed typing sequence. The accuracy rates of classifying anger, boredom, confidence, distraction, excitement, focus, frustration, happiness, hesitance, nervousness, overwhelmed, relaxation, sadness, stress, and tired, with respect to two-class modelsb, were 75% on average. The study built models by using ML algorithms and a correlation-based feature subset attribute selection method . Although the keystroke features that were used to build the model with the highest accuracy were provided, the relationship between emotion and keystroke dynamics, still, was not reported. Recently, more results related to classification on emotional data using similar feature set have been proposed. Alhothali A  reported the use of keystroke features that were extracted from arbitrarily typed keystroke sequences as reaching an 80% accuracy rate of classifying experience-sampled positive and negative emotional states. Bixler R and D’Mello S  demonstrated a 66.5% accuracy rate on average for two-class models in detecting boredom, engagement, and neutral states. The emotional data used were collected using the experience sampling method.
By applying ML methodology for building classification models from various datasets collected from different experimental setups, these studies have suggested that keystroke duration and keystroke latency can be used for model building. One therefore could hypothesize that the keystroke duration and latency may be different when subjects are in different emotional states. However, the details about the source of variance were never discussed in previous studies possibly due to the limitation of the adopted methodology. Hence, the current study aims to test the hypotheses that keystroke dynamics may be influenced by emotions.
The current study argues that the relationship between emotion and keystroke dynamics should not be too complex. By using a rigorous experiment setup, traditional statistical methods could be used to examine the variance and reveal the relationship, without the use of sophisticated ML algorithms. The study examines the variance of keystroke dynamics caused by emotions. Specifically, three hypotheses were tested. It is hypothesized that keystroke duration, keystroke latency, and the accuracy rate of a keyboard typing task are influenced by emotions. The study aims to answer two research questions. First, do the variance in the keystroke features that are ordinarily used for model building (i.e. keystroke duration, keystroke latency, accuracy rate) in previous studies exceeds significance level under different emotional states. Second, how large is the variance contributed by emotions in these keystroke features.
This study is under the research project “A study of interactions between cognition, emotion and physiology (Protocol No: 100-014-E),” which was approved by the Institution Review Board (IRB) of the National Taiwan University Hospital Hsinchu Branch. Written Informed consents were obtained from all subjects before the experiment.
Twenty-seven subjects ranging in age between 19 and 27 (M = 21.5, SD = 2.3) performed keyboard typing tasks right after presented with emotional stimuli. The subjects were college students selected from a university in Taiwan, with normal or corrected-to-normal vision and normal range of finger movement. All the subjects self-reported that they were non-smoker, healthy, with no history of brain injury and cardiovascular problems.
A number sequence was used as the target typing text instead of an alphabet sequence or symbols to avoid possible interference that can be caused by linguistic context on the subject’s emotional states. In all the various number sequences used in our pilot experiments , we found the existence of the difference in keystroke typing between the subjects in different emotional states. However, we also found that the relationship between the keystroke typing and emotional states may be different due to different keys that are typed and also the order of typing. A comparison of keystroke typing between emotional states using different number sequences may reduce the power of statistical tests (given a same number of trials). Hence, to conduct a more conservative comparison across emotion and to enhance the generalizability of this study, we decided to use a single number sequence that is designed to be general. In the current study, we designed the target typing text “24357980” to 1) be easy to type without requiring the subjects to perform abrupt changes in their posture, 2) have the number of digits fairly distributed on a number pad, and 3) encourage all of the subjects to maintain the same posture (i.e., in terms of finger usage) when typing the given sequence  (see Figure 1(b) for more detail). The time length of the experiment was designed to be as short as possible to avoid the subjects from being tired of typing on the keyboard. Note that all the subjects indeed reported that they were not fatigued after the experiment.
Stimuli and self-report
The stimuli we used were 60 pictures selected from the IAPS database, which is being developed and distributed by the NIMH Center for Emotion and Attention (CSEA) at the University of Florida . The IAPS is developed to provide a set of normative emotional stimuli for experimental investigations of emotion and attention and can be easily obtained through e-mail application. The IAPS database contains various affective pictures proved to be capable of inducing diverse emotions in the affective space . The pictures we used as the stimuli were selected from IAPS database complying the IAPS picture set selection protocol described in . The protocol includes the constraint about the number of pictures used in a single experiment, and the distribution of the emotions that are expected to be induced by the selected pictures. Stimulus order was randomized by a computer program for each subject, in order to balance the position of a particular stimulus within the series across the subjects.
The SAM is a non-verbal pictorial assessment that is designed to assess the emotional dimensions (i.e. valence and arousal) directly by means of two sets of graphical manikins. The SAM has been extensively tested in conjunction with the IAPS and has been used in diverse theoretical studies and applications [15–17]. The SAM takes a very short time to complete (5 to 10 seconds). For using the SAM, there is little chance of confusion with terms as in verbal assessments. The SAM was also reported to be capable of indexing cross-cultural results  and the results obtained using a Semantic Differential scale (the verbal scale provided in ). The SAM that we used was identical to the 9-point rating scale version of SAM that was used in , in which the SAM ranges from a smiling, happy figure to a frowning, unhappy figure when representing the affective valence dimension. On the other hand, for the arousal dimension, the SAM ranges from an excited, wide-eyed figure to a relaxed, sleepy figure. Ratings are scored such that 9 represents a high rating on each dimension (i.e. positive valence, high arousal), and 1 represents a low rating on each dimension (i.e. negative valence, low arousal).
During the experiment, a subject sat on an office chair (0.50 × 0.51 m, height 0.43 m), in a small, quiet office (7.6 × 3.2 m) without people. The office was with window and the ventilation was guaranteed. The computer system (acer Veriton M2610, processor: Intel Core i3-2120 3.3G/3M/65 W, memory: 4 GB DDR3-1066, operating system: Microsoft Windows 7 Professional 64bit) used by the subject was put under a desk (0.70 × 1.26 m, height 0.73 m). The subject was seated approximately 0.66 m from the computer screen (ViewSonic VE700, 17 inch, 1280 × 1024 in resolution). The keyboard used by the subject was an acer KU-0355 (18.2 × 45.6 cm, normal keyboard with the United States layout, typically used for Windows operating system) connected to the computer system used through USB 2.0 communication interface. The distance between the center of adjacent keys (size: 1.2 × 1.2 cm) of the number pad used was 2 cm. Keyboard lifts (the two small supports at the back of the keyboard) which will raise the back of the keyboard for 0.8 cm when used, were not used in this experiment. The subject was sat approximately 0.52 m from the center of the number pad (i.e. the digit “5” of the number pad). The software designed for keystroke collection was developed using C# project built by using Visual Studio 2008 and was executed on the .NET framework (version 3.5) platform. The reason of using C# in developing this software is that Microsoft Windows operating systems provide more sufficient Application Programming Interfaces (APIs) for C# to detect keystroke-interrupt than for other programming language such as R, Matlab, Java, and Python.
In total, 60 (trials) × 27 (subjects) = 1,620 rows of the raw data were collected during the experiment. To examine the keyboard typing patterns, single keystroke analysis  was applied to our raw data. Keystroke durations and keystroke latencies are ordinarily used in previous studies for single keystroke analysis [5, 7, 13]. The keystroke duration is the time that elapses from the key press to the key release, whereas the keystroke latency is the time that elapses from one key release to the next key press .
In our analysis, a sequence typed is a “correctly typed sequence” if the target typing text was correctly typed and “incorrectly typed sequence” if incorrectly typed. For example, if a subject typed “244357980”, in which the “4” at the 2nd digit is misplaced, such that the sequence typed is considered as an incorrectly typed sequence. A pre-processing routine was applied to the raw data to separate all the correctly typed sequences from incorrectly typed sequences.
Keystroke duration and keystroke latency features were only extracted from the correctly typed sequences (90.2% of the 1620 samples). The extracted keystroke duration and keystroke latency features were submitted to two two-way 3 (Valence: negative, neutral, and positive) × 3 (Arousal: low, medium, and high) Repeat Measures ANOVAs , respectively. To analyse the accuracy rate of keyboard typing, the accuracy data (0 for incorrectly typed sequence and 1 for correctly typed sequence) of all the typed sequences was submitted to a two-way 3 (Valence: negative, neutral, and positive) × 3 (Arousal: low, medium, and high) Repeat Measures ANOVA.
The 9-point scale SAM ratings of the valence and arousal were translated into three levels of the ANOVA factor Valence and Arousal. The significance level α of the entire statistical hypothesis tests used in this paper was set to 0.05.
Influence of emotion on keystroke duration
Descriptive statistics of keystroke duration under independent variables valence × arousal
95% confidence interval
Repeated measures two-way ANOVA table for keystroke duration
Source of variance
Valence × Arousal***
Error (Valence × Arousal)
Influence of emotion on keystroke latency
Descriptive statistics of keystroke latency under independent variables valence × arousal
95% confidence interval
Repeated measures two-way ANOVA table for keystroke latency
Source of variance
Valence × Arousal***
Error (Valence × Arousal)
Influence of emotion on accuracy rate
Descriptive statistics of accuracy rate under independent variables valence × arousal
95% confidence interval
Repeated measures two-way ANOVA table for accuracy rate
Source of variance
Valence × Arousal***
Error (Valence × Arousal)
Prior studies have highlighted the possibility of using keyboard typing data to detect emotions. Specifically, keystroke duration, keystroke latency, and accuracy rate of keyboard typing were used as input features for model building. These results have led to three hypothesized relationships. The relationship between keystroke duration and emotions, the relationship between keystroke latency and emotions, and the relationship between accuracy rate of keyboard typing and emotions. Hence, the current study tests these three hypothesized relationships. The results of our experiment using the fix target typing text and the 60 stimuli selected from the IAPS database support the hypothesis that the keystroke duration, keystroke latency, and also the accuracy rate of typing, are influenced by emotional states. The results further indicate that the keystroke duration is more sensitive to Valence, whereas the accuracy rate is more sensitive to Arousal. Moreover, the keystroke latency is affected by both Valence and Arousal, with these two variables interacts with each other.
It is worth to note that the size of the emotional effects that were found is small (see Tables 2, 4, and 6), compare to the individual variability. The finding suggests that although previous studies have built intelligent systems that act user-independently in detecting emotional states of users based on the keystroke dynamics, the accuracy rate of the detection could be further improved if personalized models are utilized. In addition, this finding also provides an explanation to that why real-world applications which authenticate the identity of users by monitoring keystrokes may not be interfered by the emotional states of users.
The research questions about the three hypothesized relationships between emotions and keystroke dynamics are answered by using traditional statistical methods instead of sophisticated ML algorithms. The source of variance was examined and the emotional factors (in terms of valence and arousal) that affect keystroke duration, keystroke latency, and the accuracy rate of keyboard typing, were identified. To summarize, the evidence that were found supports all the three hypotheses.
aA model that is readable means that the model is described clearly, with the relationship between independent variable and dependent variable identified and could be easily interpreted.
bA two-class model is the type of classification model that classifies instances into two classes (i.e. is an instance with the target label or not with the target label).
This work was fully supported by the Taiwan Ministry of Science and Technology under grant numbers NSC-102-2220-E-009-023 and NSC-102-2627-E-010-001. This work was also supported in part by the UST-UCSD International Center of Excellence in Advanced Bioengineering sponsored by the Taiwan Ministry of Science and Technology I-RiCE Program under grant number NSC-101-2911-I-009-101; and in part by “Aim for the Top University Plan” of the National Chiao Tung University and Ministry of Education, Taiwan, R.O.C.
- Picard RW: Affective Computing. Cambridge MA: The MIT Press; 2000.Google Scholar
- Cohen I, Sebe N, Garg A, Chen LS, Huang TS: Facial expression recognition from video sequences: temporal and static modeling. Comput Vis Image Underst 2003, 91: 160–187. 10.1016/S1077-3142(03)00081-XView ArticleGoogle Scholar
- Cowie R, Douglas-Cowie E, Tsapatsoulis N, Votsis G, Kollias S, Fellenz W, Taylor JG: Emotion recognition in human-computer interaction. IEEE Signal Process Mag 2001, 18: 32–80. 10.1109/79.911197View ArticleGoogle Scholar
- Kim KH, Bang SW, Kim SR: Emotion recognition system using short-term monitoring of physiological signals. Med Biol Eng Comput 2004, 42: 419–427. 10.1007/BF02344719View ArticleGoogle Scholar
- Zimmermann P, Guttormsen S, Danuser B, Gomez P: Affective computing–a rationale for measuring mood with mouse and keyboard. Int J Occup Saf Ergon 2003, 9: 539–551.Google Scholar
- Vizer LM, Zhou L, Sears A: Automated stress detection using keystroke and linguistic features: an exploratory study. Int J Hum Comput Stud 2009, 67: 870–886. 10.1016/j.ijhcs.2009.07.005View ArticleGoogle Scholar
- Epp C, Lippold M, Mandryk RL: Identifying Emotional States Using Keystroke Dynamics. In Proceedings of the 2011 Annual Conference on Human Factors in Computing Systems. Vancouver, BC, Canada: ACM; 2011:715–724.View ArticleGoogle Scholar
- Hall MA: Correlation-Based Feature Selection for Machine Learning. Hamilton, New Zealand: The University of Waikato; 1999.Google Scholar
- Alhothali A: Modeling User Affect Using Interaction Events. Waterloo, Ontario, Canada: University of Waterloo, School of Computer Science; 2011.Google Scholar
- Bixler R, D’Mello S: Detecting Boredom and Engagement During Writing With Keystroke Analysis, Task Appraisals, and Stable Traits. In Proceedings of the 2013 International Conference on Intelligent user Interfaces. Santa Monica, California, USA: ACM; 2013:225–234.View ArticleGoogle Scholar
- Lang PJ, Bradley MM, Cuthbert BN: International Affective Picture System (IAPS): Affective Ratings of Pictures and Instruction Manual. Gainesville, FL: University of Florida; 2008.Google Scholar
- Lang PJ: Behavioral Treatment and Bio-Behavioral Assessment: Computer Applications. In Technology in Mental Health Care Delivery Systems. Edited by: Sidowski J, Johnson J, Williams T. Norwood, NJ: Ablex Pub. Corp; 1980:119–137.Google Scholar
- Tsui WH, Lee PM, Hsiao TC: The Effect of Emotion on Keystroke: An Experimental Study Using Facial Feedback Hypothesis. 35th Annual International Conference of the IEEE Engineering in Medicine and Biology Society (EMBC’13); July 3–7, 2013; Osaka, Japan 2013, 2870–2873.View ArticleGoogle Scholar
- Bradley MM, Lang PJ: Emotion and Motivation. In Handbook of Psychophysiology. 3rd edition. Edited by: Cacioppo JT, Tassinary LG, Berntson G. New York, NY: Cambridge University Press; 2007:581–607.View ArticleGoogle Scholar
- Bradley MM: Emotional Memory: A Dimensional Analysis. In Emotions: Essays on Emotion Theory. Edited by: Goozen SHM, Poll NE, Sergeant JA. Hillsdale, NJ: Lawrence Erlbaum; 1994:97–134.Google Scholar
- Bolls PD, Lang A, Potter RF: The effects of message valence and listener arousal on attention, memory, and facial muscular responses to radio advertisements. Comm Res 2001, 28: 627–651. 10.1177/009365001028005003View ArticleGoogle Scholar
- Chang C: The impacts of emotion elicited by print political advertising on candidate evaluation. Media Psychol 2001, 3: 91–118. 10.1207/S1532785XMEP0302_01View ArticleGoogle Scholar
- Morris JD: Observations: SAM: the Self-Assessment Manikin; an efficient cross-cultural measurement of emotional response. J Advert Res 1995, 35: 63–68.Google Scholar
- Mehrabian A, Russell JA: An Approach to Environmental Psychology. Cambridge: MA: the MIT Press; 1974.Google Scholar
- Bergadano F, Gunetti D, Picardi C: Identity verification through dynamic keystroke analysis. Intell Data Anal 2003, 7: 469–496.Google Scholar
- Monrose F, Rubin AD: Keystroke dynamics as a biometric for authentication. Future Generat Comput Syst 2000, 16: 351–359. 10.1016/S0167-739X(99)00059-XView ArticleGoogle Scholar
- Langsrud Ø: ANOVA for unbalanced data: use type II instead of type III sums of squares. Stat Comput 2003, 13: 163–167. 10.1023/A:1023260610025MathSciNetView ArticleGoogle Scholar
- Bradley MM, Lang PJ: Measuring emotion: the self-assessment manikin and the semantic differential. J Behav Ther Exp Psychiatr 1994, 25: 49–59. 10.1016/0005-7916(94)90063-9View ArticleGoogle Scholar
- Lee PM, Teng Y, Hsiao TC: XCSF for Prediction on Emotion Induced by Image Based on Dimensional Theory of Emotion. In Proceedings of the Fourteenth International Conference on Genetic and Evolutionary Computation Conference Companion. Philadelphia, USA: ACM; 2012:375–382.View ArticleGoogle Scholar
This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/4.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly credited. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.