Matthew G. Wisniewski

Learning-Related Shifts in Generalization Gradients for Complex Sounds

Learning to discriminate stimuli can alter how one distinguishes related stimuli. For instance, training an individual to differentiate between two stimuli along a single dimension can alter how that individual generalizes learned responses. In this study, we examined the persistence of shifts in generalization gradients after training with sounds. University students were trained to differentiate two sounds that varied along a complex acoustic dimension. The students were subsequently tested on their ability to recognize a sound that they had experienced during training when it was presented among several novel sounds varying along this same dimension. Peak shift was observed in Experiment 1, in which generalization tests immediately followed training, and in Experiment 2, in which the tests were delayed by 24 h. These findings further support the universality of generalization processes across species, modalities, and levels of stimulus complexity. They also raise new questions about the mechanisms underlying learning-related shifts in generalization gradients. The sound stimuli from this study are available as .wav files from http://lb.psychonomic-journals.org/content/supplemental.

Temporal dynamics in auditory perceptual learning: impact of sequencing and incidental learning.

Training can improve perceptual sensitivities. We examined whether the temporal dynamics and the incidental versus intentional nature of training are important. Within the context of a birdsong rate discrimination task, we examined whether the sequencing of pretesting exposure to the stimuli mattered. Easy-to-hard (progressive) sequencing of stimuli during preexposure led to a more accurate performance with the critical difficult contrast and greater generalization to new contrasts in the task, compared with equally variable training in either a random or an antiprogressive order. This greater accuracy was also evident when participants experienced the progressively sequenced stimuli in a different incidental learning task that did not involve direct auditory training. The results clearly show the importance of temporal dynamics (sequencing) in learning and show that the progressive training advantages cannot be fully explained by direct associations between stimulus features and the corresponding responses. The current findings are consistent with a hierarchical account of perceptual learning, among other possibilities, but not with explanations that focus on stimulus variability.

Familiarity with Speech Affects Cortical Processing of Auditory Distance Cues and Increases Acuity

Several acoustic cues contribute to auditory distance estimation. Nonacoustic cues, including familiarity, may also play a role. We tested participants’ ability to distinguish the distances of acoustically similar sounds that differed in familiarity. Participants were better able to judge the distances of familiar sounds. Electroencephalographic (EEG) recordings collected while participants performed this auditory distance judgment task revealed that several cortical regions responded in different ways depending on sound familiarity. Surprisingly, these differences were observed in auditory cortical regions as well as other cortical regions distributed throughout both hemispheres. These data suggest that learning about subtle, distance-dependent variations in complex speech sounds involves processing in a broad cortical network that contributes both to speech recognition and to how spatial information is extracted from speech.

Temporal dynamics of generalization and representational distortion.

Individuals generalize differently depending on the extent of their past experiences and what they learn from them. For instance, the peak of generalization can shift from a familiar stimulus to novel ones when the familiar stimulus has been repeatedly discriminated from another, similar stimulus. We examined how the amount of experience students had in discriminating complex sounds impacted their later generalization. As training increased, participants improved at distinguishing sounds, but their tendency to respond to certain novel sounds increased. With additional training, however, this shift dissipated. The results suggest that shifts in generalization may correspond to transitional states of perceptual learning and that learning-related changes in perceptual sensitivities involve more than just incremental increases in feature selectivity.

Predicting shifts in generalization gradients with perceptrons

Perceptron models have been used extensively to model perceptual learning and the effects of discrimination training on generalization, as well as to explore natural classification mechanisms. Here, we assess the ability of existing models to account for the time course of generalization shifts that occur when individuals learn to distinguish sounds. A set of simulations demonstrates that commonly used single-layer and multilayer perceptron networks do not predict transitory shifts in generalization over the course of training but that such dynamics can be accounted for when the output functions of these networks are modified to mimic the properties of cortical tuning curves. The simulations further suggest that prudent selection of stimuli and training criteria can allow for more precise predictions of learning-related shifts in generalization gradients in behavioral experiments. In particular, the simulations predict that individuals will show maximal peak shift after different numbers of trials, that easier contrasts will lead to slower development of shifted peaks, and that whether generalization shifts persist or dissipate will depend on which stimulus dimensions individuals use to distinguish stimuli and how those dimensions are neurally encoded.

Learning to discriminate frequency modulation rate can benefit and worsen pitch acuity

Participants were trained to discriminate frequency modulation rates (FM-rate training) or Gabor patch orientations (visual training) in a same-different task for two different training lengths. Test discriminations involved trains of FM sweeps with identical modulation rates, but different frequencies. FM-rate training enhanced test accuracy (relative to visual) when sweep trains contained frequencies similar to training. For extended FM-rate training, the opposite was true for trains shifted one octave higher. In contrast to previous work, generalization of learning to the untrained dimension (pitch) was not well accounted for by conceptual learning. Mechanisms of stimulus learning may better explain the current cross-dimensional generalization.

Benefits of fading in perceptual learning are driven by more than dimensional attention

Individuals learn to classify percepts effectively when the task is initially easy and then gradually increases in difficulty. Some suggest that this is because easy-to-discriminate events help learners focus attention on discrimination-relevant dimensions. Here, we tested whether such attentional-spotlighting accounts are sufficient to explain easy-to-hard effects in auditory perceptual learning. In two experiments, participants were trained to discriminate periodic, frequency-modulated (FM) tones in two separate frequency ranges (300–600 Hz or 3000–6000 Hz). In one frequency range, sounds gradually increased in similarity as training progressed. In the other, stimulus similarity was constant throughout training. After training, participants showed better performance in their progressively trained frequency range, even though the discrimination-relevant dimension across ranges was the same. Learning theories that posit experience-dependent changes in stimulus representations and/or the strengthening of associations with differential responses, predict the observed specificity of easy-to-hard effects, whereas attentional-spotlighting theories do not. Calibrating the difficulty and temporal sequencing of training experiences to support more incremental representation-based learning can enhance the effectiveness of practice beyond any benefits gained from explicitly highlighting relevant dimensions.

Learning-related improvements in auditory detection sensitivities correlate with changes in sensory- and decision-related components of the event-related potential

Listener performance in an auditory detection task can improve with practice (Zwislocki, Maire, Feldman, & Rubin, 1958). This could result from a selective attention process and/or sensory plasticity (e.g., if trained stimuli receive increased cortical representation). Here, listeners were trained to detect either an 861-Hz or 1058-Hz tone (counterbalanced across participants) presented in a noise masker. On the following day, high-density EEG was collected while listeners: 1) attempted to detect 861-Hz and 1058-Hz tones in noise at an SNR of -21 dB, and 2) passively heard the same tones presented in quiet. Listeners were significantly better at detecting tones at their trained frequency. In addition, P3 amplitudes were larger for trained than for untrained tones during the detection task. During passive exposure to the same tones, P2 amplitudes were similarly larger for trained than for untrained tones. The difference in P3 amplitudes suggests that training leads to more efficient decisional processing, ...

Chickadee songs provide hidden clues to singers’ locations

Funding: LMG was supported by an Izaak Walton Killam Memorial Scholarship (IWKMS) at UofA and is currently a BBSRC Anniversary Future Leader Fellow.

Learning-related improvements in auditory detection sensitivities correlate with neural changes observable during active and passive sound processing

Auditory detection can improve with training. Though the majority of theorists posit that such improvements relate to dynamics of selective attention, longer-term sensory plasticity may also play a role (e.g., increased cortical processing for trained sounds). Here, listeners were trained in a single session to detect either an 861-Hz or 1058-Hz tone (counterbalanced across participants) presented in a frozen white noise masker. On the following day, EEG was collected while listeners: 1) attempted to detect 861-Hz and 1058-Hz tones at an SNR of -21 dB, and 2) passively heard the same tones presented in quiet. Listeners were significantly better at detecting tones at their trained frequency. In addition, P3 amplitudes were larger for trained than for untrained tones during the active detection task. During passive exposure to the same tones, P2 amplitudes were similarly larger for trained than for untrained tones. The difference in P3 amplitudes suggests that training leads to more efficient post-sensory p...