Sequential Grouping

Question 1

Van Noorden (1975)

Answer 1

• evidence for factors influencing stream segregation:
  > frequency separation
  > rate of presentation
  > listening set/instructions
• presented participants with simple A-B-A sequence
• when frequency separation was smaller, rate required to be faster for streams to segregate compared to when frequency separation is bigger
• temporal coherence boundary (TCB; hearing 2 streams) was highly dependant on rate of presentation
• fission boundary (FB; hearing single stream) was less dependant on rate and more dependant on frequency separation
• in area between the two boundaries, stream segregation is based on instructions, as it is possible to flip between the two
  (similar to principle of exclusive allocation - can flip between two but it is not possible to hear both one and two streams at the same time - must be allocated to one of the two)

Question 2

Bregaman (1978) - Build up & Decay

Answer 2

• investigated cumulative effects of stream segregation
• tendency to hear two streams builds up over time
• used A-B-A sequence to measure splitting threshold between hearing one or two streams based on repetition of sequence in a ‘package’
• the more repetitions of a sequence, the more likely it is to segregate in to two streams
• if silent interval is introduced for longer than 4sec, build-up is lost, if shorter than 4sec, it continues

Question 3

Anistis & Saida (1985) - Resetting

Answer 3

• found that build-up process continued for about 30 sec and broke when there was an abrupt change in the sequence
• frequency separation of more than 2 semi-tones reset build up, if less than 2 semitones, build-up continued

Question 4

Bregman & Rogers (1988) - Resetting

Answer 4

• found that resetting is not specific to changes in frequency
• introduced changes to ITD
• test sequence was preceded by inducer sequence of same frequency, but changed spatial location of inducer
• with no inducer, segregation boundary was about 13 semitones (tolerated big difference before segregation), and with an inducer with no differences with test sequence, segregation boundary was smaller (8 semitones, tolerated less difference before segregation) due to effects of buildup
• however, when sudden change in ITD was introduced, segregation boundary was about 11 semitones, indicating build up had been reset as a bigger difference in frequency was required to hear segregation
• interestingly, if ILD’s were also introduced, there was an asymmetric relationship with build up: so if inducer > test was soft > loud, build up was lost, but not vice versa

Question 5

Carylon et al (2001) - Role of attention

Answer 5

• compared build up of A-B-A sequence under 3 conditions:
  > no sounds in right ear
  > ignore sounds in right ear (distractor)
  > attend to sounds in both
• in baseline, segregation was based on frequency separation, and same for distractor task
• in last condition, there was no build up which suggested that attention was required for build up of streaming

Question 6

Sussman et al (1999) - MMN

Answer 6

• investigated role of attention in build up using EEG study of MMN where deviant tones show up on EEG recording
• deviant tone was able to be detected in conditions where streams had segregated (baed on differences in frequency), but not if it was heard as a single tone
• participants were played sequences when engaged in other task (reading a book) and found MMN response to deviant tones even when not attending
• this would suggest that attention not necessary for build up therefore Carylon et al (2001) can be interpreted as loss of build up due to abrupt change in attention to spatial location, as opposed to due to attention entirely

Question 7

Van Noorden (1975) - Timbre of Complex Tones

Answer 7

• investigated influence of timbre of complex tones on stream segregation
• alternated complex tones with F0 missing, with pure tones corresponding to F0 of complex tone which introduced timbre differences due to changes in the spectral shape
• found that changes in timbre led to segregation despite pitch being the same

Question 8

Wessel Illusion (1979) - Timbre of Complex Tones

Answer 8

• investigated influence of pitch on segregation if complex tones
• alternately removed high frequencies (for even harmonics) and low frequencies (for odd harmonics) introducing timbre differences
• this introduced segregation of high and low streams, also changing direction of pattern from ascending to descending

Question 9

Bregman, Lao & Levitan (1990) - Competition between pitch and timbre

Answer 9

• competition between pitch and timbre influences streaming
• grouping depends on which of the two factors have a bigger difference between tones i.e. bigger difference in pitch than timbre will lead to grouping by pitch and vice versa

Question 10

Hartmann & Johnson (1991) - PCH

Answer 10

• PCH suggests that stream segregation doesn’t occur unless there are differences in the way auditory information is represented in the brain/excites different neural channels
• ear preferentially responds to different frequencies hitting different areas of basilar membrane (lower frequencies at apex, higher at base)
• PCH suggests that if excitation patterns are the same, segregation won’t occur

Question 11

Singh & Bregman (1997)

Answer 11

• changed temporal envelope of tones by time-reversing it changing timbre but not spectral content of frequency. This simply changed attack and decay pattern
• found that this led to additional streaming segregation
• HOWEVER, cannot be concluded that changes int temporal envelope alone led to segregation because methodology in this study also changed F0 which introduce PC cues

Question 12

Vliegen et al (1999)

Answer 12

• demonstrated that using unresolved harmonics within same frequency band do not introduce additional PC cues, allowing PCH to be tested
• found that stream segregation can occur based on differences in F0 despite no differences in excitation pattern
• Roberts, Glasberg & Moore (2002) - used similar stimuli and found that this effect can be strong
• concluding that sequential grouping is based on Gestalt principles of similarity as opposed of PC cues

Question 13

Darwin & Hukin (1999) - Spatial Location

Answer 13

• used dichotic presentation to investigated ITD’s
• natural speech was processed so that each sentence was monotonised (on a constant F0)
• listeners presented sentences simultaneously in both ears and asked to detect target words
• found that they tended to track by localisation as opposed to common F0

Question 14

Drone Illusion (1979)

Answer 14

• if melody is scattered between ears, then streams are segregated based on location
• however if a drone note of the same frequency repeats at the same as melody in opposite ear, the melody becomes recognisable
• drone weakens spatial location cues, pulling melody towards centre of space, as grouping then occurs based on similarity

Question 15

Deutsch (1975) - Scale Illusion

Answer 15

• listeners were simultaneously presented with two scales (ascending and descending), except every other note swapped between ears
• found that listeners did not group between ear of arrival but by frequency proximity (all high notes heard in one ear and low in other)
• were unable to identify which ear notes came from
• highlights that spatial location can be quite a weak grouping cue and can lose out to to other cues e.g. proximity