6.3 ASSR and CAEP: frequency-specific and cortical responses

The ABR records a transient response: a brief stimulus produces a brief response, and we average across many stimulus presentations. Two younger members of the evoked-potential family work differently.

The auditory steady-state response (ASSR) uses a continuous amplitude- or frequency-modulated tone and looks not for transient peaks but for a steady periodic component of the EEG locked to the modulation frequency. Because multiple modulated tones at different carriers and different modulation frequencies can be presented simultaneously and analysed independently in the frequency domain, the ASSR enables true parallel multi-frequency objective threshold estimation — a major efficiency gain for pediatric testing.

The cortical auditory evoked potential (CAEP) records the brain’s response over a much longer post-stimulus window (50–500 ms) and isolates cortical rather than brainstem generators. CAEP latencies are slower, amplitudes larger, and the responses index higher-order processing — including the developmental status of the auditory cortex itself.

This lesson covers both, and closes the chapter.

ASSR — the frequency-domain view

The ASSR setup is identical to ABR — same electrodes, same shielded booth, same averaging hardware — but the stimulus is different. Instead of a click, deliver a continuous tone whose amplitude is modulated at a low frequency, typically 80–100 Hz in adults or 40 Hz in older children. The brain, processing this stimulus, produces an EEG component at the modulation frequency. A 1000 Hz carrier amplitude-modulated at 91 Hz, presented for a few minutes, will produce a 91 Hz EEG component whose amplitude grows with carrier intensity until carrier threshold is reached.

The clinical question is not “is the EEG ringing at 91 Hz” — there’s always some 91 Hz energy in the EEG — but is the EEG at 91 Hz phase-locked to the modulator? Phase-locking is tested with the Hotelling $T^2$ statistic applied to the complex Fourier coefficient of the EEG at the modulation frequency over many short windows. Hotelling $T^2$ tests the null hypothesis that the complex coefficient has zero mean — i.e., that the phase is randomly distributed. If $p < 0.05$ across enough windows, the response is declared present.

Why multi-frequency parallel testing works

The genius of the ASSR is frequency-domain orthogonality: present multiple carriers at different audiometric frequencies (500, 1000, 2000, 4000 Hz), each modulated at a different modulation frequency (77, 85, 93, 101 Hz), all simultaneously. Each carrier produces a phase-locked EEG response only at its modulation frequency. A single Fourier analysis of the EEG separates the four responses cleanly — same recording time, four frequencies tested. Add the other ear, with four different modulation frequencies, and you have an 8-carrier, 8-modulation-frequency, single-recording protocol that probes both ears at four frequencies at once.

A multi-frequency ASSR battery for pediatric threshold estimation takes 30–60 minutes for both ears at four frequencies — competitive with tone-burst ABR but yielding more information per minute of recording.

ASSR vs ABR — when to use which

ABR remains the gold standard for neurodiagnostic questions (retrocochlear, brainstem); ASSR is increasingly preferred for threshold-only pediatric objective measurement. Modern instruments commonly perform both in the same session.

CAEP — the cortical response

The cortical auditory evoked potential is the long-latency response that follows the brainstem response. The standard CAEP, recorded from Cz with mastoid reference and bandpass-filtered 1–30 Hz, contains:

• P1 at ~50 ms (~3 µV in adults, ~7 µV in infants; the dominant infant response).
• N1 at 100 ms (−4 µV in adults; small in infants under 18 months).
• P2 at ~180 ms (~3 µV).
• N2 at ~250 ms.

These late components index cortical processing — the response is generated by primary and secondary auditory cortices and their associated areas. CAEP amplitudes are larger than ABR amplitudes (microvolts rather than tenths of microvolts), which makes them less averaging-intensive, but the longer latency and lower stimulus rate (typically 0.5–1 Hz, since each response needs ~500 ms to complete) makes the protocol slower overall.

CAEP has two distinct clinical uses:

1. Cochlear implant outcome assessment

A young child receiving a cochlear implant has, by definition, congenital profound hearing loss. Whether the implant is working — that is, whether the implant’s electrical stimulation is producing usable cortical activation — cannot be assessed behaviourally in a 12-month-old. CAEP fills this gap: an acoustic stimulus presented through a loudspeaker, or a direct electrical stimulus through the implant’s externally-controlled processor, should produce a P1-N1-P2 response if the auditory cortex is being activated. The presence or absence of the P1 in particular has become a standard biomarker for cortical activation in pediatric CI populations.

P1 latency in children is age-dependent (longer in younger children, shortening to adult values around age 8) and developmentally informative. Sharma et al. (2002, 2005, 2007) established that children who receive cochlear implants before age 3.5 show P1 latencies that normalise to age-appropriate values within 6–8 months of activation; children implanted after age 7 typically retain abnormal P1 latencies despite the implant. This “sensitive period” has become one of the strongest justifications for early implantation in eligible candidates.

2. Central auditory processing assessment

Patients with normal peripheral hearing but unexplained listening difficulties (children diagnosed with “central auditory processing disorder”, adults with post-concussion auditory complaints, individuals with developmental language disorder) may show subtle CAEP abnormalities — delayed N1 latencies, reduced P2 amplitudes, atypical mismatch negativity (MMN, a 150–250 ms negativity to a deviant stimulus in an oddball sequence). The clinical utility of these findings remains contested — they correlate with behavioural deficits in groups but are not diagnostic in individuals — but CAEP remains a research and clinical tool for the population of patients whose audiograms and middle-ear status are unremarkable but whose hearing function is impaired.

Closing the chapter

That closes Chapter 6. The arc: signal averaging — a measurement principle borrowed from radar — recovers a 0.3 µV time-locked response from a 1.5 µV EEG floor. The ABR records the brainstem’s five-stage activation in the first 10 ms after a click; its waveform supports both threshold estimation and neurodiagnostic localisation. The ASSR exploits frequency-domain orthogonality to test multiple frequencies in parallel — a major efficiency gain that has largely replaced tone-burst ABR for routine pediatric threshold measurement. The CAEP probes cortical activation at longer latencies and is the standard biomarker for cochlear-implant outcomes in pre-lingual deaf children.

Together with otoacoustic emissions (Ch 5), evoked potentials complete the audiologist’s objective testing toolkit. We have now finished the assessment chapters of the book. The remaining four chapters cover the audiologist’s interventions — hearing aids, real-ear verification, cochlear implants, and bone-conduction devices.

Next chapter: Ch 7 — Hearing aids. The DSP pipeline, wide-dynamic-range compression, directional processing, and feedback cancellation — the modern hearing aid as a small but extraordinarily sophisticated computer attached to the ear.