Publications

Abstract

Humans are nature’s most intelligent and prolific users of external props and aids (such as written texts, slide-rules and software packages). Here we introduce a method for investigating how people make active use of their task environment during problem-solving and apply this approach to the non-verbal Raven Advanced Progressive Matrices test for fluid intelligence. We designed a click-and-drag version of the Raven test in which participants could create different external spatial configurations while solving the puzzles. In our first study, we observed that the click-and-drag test was better than the conventional static test at predicting academic achievement of university students. This pattern of results was partially replicated in a novel sample. Importantly, environment-altering actions were clustered in between periods of apparent inactivity, suggesting that problem-solvers were delicately balancing the execution of internal and external cognitive operations. We observed a systematic relationship between this critical phasic temporal signature and improved test performance. Our approach is widely applicable and offers an opportunity to quantitatively assess a powerful, although understudied, feature of human intelligence: our ability to use external objects, props and aids to solve complex problems.

Abstract

In a Turing test, a judge decides whether their conversation partner is either a machine or human. What cues does the judge use to determine this? In particular, are presumably unique features of human language actually perceived as humanlike? Participants rated the humanness of a set of sentences that were manipulated for grammatical construction: linear right-branching or hierarchical center-embedded and their plausibility with regard to world knowledge.

We found that center-embedded sentences are perceived as less humanlike than right-branching sentences and more plausible sentences are regarded as more humanlike. However, the effect of plausibility of the sentence on perceived humanness is smaller for center-embedded sentences than for right-branching sentences.

Participants also rated a conversation with either correct or incorrect use of the context by the agent. No effect of context use was found. Also, participants rated a full transcript of either a real human or a real chatbot, and we found that chatbots were reliably perceived as less humanlike than real humans, in line with our expectation. We did, however, find individual differences between chatbots and humans.

Abstract

There is mounting evidence that the ease of producing and understanding language depends not only on the frequencies of individual words but also on the frequencies of word combinations. However, in two picture description experiments, Janssen and Barber (2012) found that French and Spanish speakers' speech onset latencies for short phrases depended exclusively on the frequencies of the phrases but not on the frequencies of the individual words. They suggested that speakers retrieved phrase-sized units from the mental lexicon. In the present study, we examined whether the time required to plan complex noun phrases in Dutch would likewise depend only on phrase frequencies. Participants described line drawings in phrases such as rode schoen [red shoe] (Experiments 1 and 2) or de rode schoen [the red shoe] (Experiment 3). Replicating Janssen and Barber's findings, utterance onset latencies depended on the frequencies of the phrases but, deviating from their findings, also depended on the frequencies of the adjectives in adjective-noun phrases and the frequencies of the nouns in determiner-adjective-noun phrases. We conclude that individual word frequencies and phrase frequencies both affect the time needed to produce noun phrases and discuss how these findings may be captured in models of the mental lexicon and of phrase production

Abstract

Background

Transcutaneous vagus nerve stimulation (tVNS) is a new, non-invasive technique being investigated as an intervention for a variety of clinical disorders, including epilepsy and depression. It is thought to exert its therapeutic effect by increasing central norepinephrine (NE) activity, but the evidence supporting this notion is limited.
Objective

In order to test for an impact of tVNS on psychophysiological and hormonal indices of noradrenergic function, we applied tVNS in concert with assessment of salivary alpha amylase (SAA) and cortisol, pupil size, and electroencephalograph (EEG) recordings.
Methods

Across three experiments, we applied real and sham tVNS to 61 healthy participants while they performed a set of simple stimulus-discrimination tasks. Before and after the task, as well as during one break, participants provided saliva samples and had their pupil size recorded. EEG was recorded throughout the task. The target for tVNS was the cymba conchae, which is heavily innervated by the auricular branch of the vagus nerve. Sham stimulation was applied to the ear lobe.
Results

P3 amplitude was not affected by tVNS (Experiment 1A: N=24; Experiment 1B: N=20; Bayes factor supporting null model=4.53), nor was pupil size (Experiment 2: N=16; interaction of treatment and time: p=0.79). However, tVNS increased SAA (Experiments 1A and 2: N=25) and attenuated the decline of salivary cortisol compared to sham (Experiment 2: N=17), as indicated by significant interactions involving treatment and time (p=.023 and p=.040, respectively).
Conclusion

These findings suggest that tVNS modulates hormonal indices but not psychophysiological indices of noradrenergic function.

Abstract

Hierarchical centre-embedded structures pose a large difficulty for language learners due to their complexity. A recent artificial grammar learning study (Lai & Poletiek, 2011) demonstrated a starting-small (SS) effect, i.e., staged-input and sufficient exposure to 0-level-of-embedding exemplars were the critical conditions in learning AnBn structures. The current study aims to test: (1) a more sophisticated type of SS (a gradually rather than discretely growing input), and (2) the frequency distribution of the input. The results indicate that SS optimally works under other conditional cues, such as a skewed frequency distribution with simple stimuli being more numerous than complex ones.

Abstract

Previous research suggests that lie detection can be improved by asking the interviewee unexpected questions. The present experiment investigates the effect of two types of unexpected questions: background questions and detail questions, on detecting lies about topics with which the interviewee is (a) familiar or (b) unfamiliar. In this experiment, 66 participants read interviews in which interviewees answered background or detail questions, either truthfully or deceptively. Those who answered deceptively could be lying about a topic they were familiar with or about a topic they were unfamiliar with. The participants were asked to judge whether the interviewees were lying. The results revealed that background questions distinguished truths from both types of lies, while the detail questions distinguished truths from unfamiliar lies, but not from familiar lies. The implications of these findings are discussed.